Archive for category: AI, Legal Tech and Family Justice

You are here: / / / / AI, Legal Tech and Family Justice

Before You Press Submit: Why AI Prompt Governance Matters in Legal Work and Family Justice

/0 Comments/in , , , /by

Artificial intelligence can help people organise evidence, draft documents and understand complex information. But the risk does not begin only with the answer an AI system produces. It begins with the prompt: the information entered, the assumptions built into the instruction and the decision to trust the output. Inspired by solicitor Genevieve Cripps’ practical work on prompt governance, this article examines what responsible AI use should look like in legal work, family proceedings and case preparation.

One of the many good points that stood out to me from Genevieve’s work

“The prompt is not merely a request for assistance. It is an information transfer.”

The AI Prompt Governance Checklist referenced in this article was developed by Genevieve Cripps. The analysis and application to family proceedings are the author’s own.

JSH Law | Legal AI, Evidence and Access to Justice

Before You Press Submit: Why Prompt Governance Matters in Legal Work, Family Justice and the Courts

The most serious risk in everyday artificial intelligence use may not begin with the model itself. It may begin with the information a person types into the prompt, the assumptions built into the instruction and the decision to trust the resulting answer without proper scrutiny.

Credit and inspiration

This article was inspired by the work of Genevieve Cripps, a solicitor whose professional interests include commercial litigation, data protection, personal data breach response, AI governance, responsible AI and emerging technology regulation.

Genevieve recently published an excellent practical resource entitled AI Prompt Governance Checklist: Before You Press Submit. Her central point is simple but important: organisations need clear, repeatable checks governing what users place into AI systems, how outputs are reviewed and when additional scrutiny is required.

You can view Genevieve’s professional work on LinkedIn here: Genevieve Cripps on LinkedIn .

In this article

  • What prompt governance actually means.
  • Why prompts can create confidentiality, privacy and evidential risks.
  • How the issue applies to solicitors, legal support providers and litigants in person.
  • Why family court material requires particular care.
  • A practical “before you press submit” framework.
  • What responsible human oversight should look like in reality.

The overlooked layer of AI risk

Much of the public debate about generative AI focuses on the model: whether it hallucinates, whether it is biased, whether it has been trained lawfully and whether its answers are reliable.

Those are legitimate concerns. But there is another layer of risk that is much closer to home.

Every day, users paste information into AI systems. They enter names, allegations, medical information, business strategies, client communications, court documents, witness evidence, financial figures and private family histories. They may do so without knowing where that information goes, whether it is retained, who can access it or whether it may be used for further model development.

The prompt is not merely a request for assistance. It is an information transfer.

Before asking whether an AI answer is useful, the user should ask whether the information should have been entered into that system at all.

What is prompt governance?

Prompt governance is the system of rules, safeguards and review processes governing how people interact with artificial intelligence.

It includes questions such as:

  • Which AI tools are approved for use?
  • What information may be entered?
  • What information must be removed, anonymised or withheld?
  • Who may create or approve important prompts?
  • How should prompts and outputs be recorded?
  • Who is responsible for checking the result?
  • When is specialist legal, privacy, security or safeguarding review required?
  • When should AI not be used at all?

Genevieve Cripps’ checklist divides this into five practical questions:

  1. Is AI appropriate for this task?
  2. Am I sharing the right information?
  3. Is my prompt clear?
  4. Can the AI output be trusted?
  5. Could the prompt or output create risk?

This is precisely the kind of implementation-focused thinking that responsible AI adoption needs. Governance cannot remain trapped inside policy papers, board presentations and abstract ethical principles. It must reach the moment when an individual user is about to press “submit”.

1. Is AI appropriate for this task?

Not every task should be delegated to an AI system.

Artificial intelligence may be useful for organising information, identifying themes, producing a first draft, simplifying language, creating a checklist or suggesting questions for further investigation.

It is far more dangerous when it is asked to make, or effectively determine, a decision requiring human judgment.

Examples include:

  • deciding whether a child is telling the truth;
  • assessing whether domestic abuse has occurred;
  • determining whether contact is safe;
  • assessing litigation capacity;
  • deciding whether allegations are credible;
  • determining whether a person poses a safeguarding risk;
  • predicting how a judge will decide a case;
  • deciding whether evidence should be reported to the police or a local authority.

AI may assist a person to identify relevant questions. It should not replace the careful, accountable and context-sensitive judgment required in high-impact legal and safeguarding decisions.

Family justice warning

A child arrangements case is not a neutral document-processing exercise. It may involve domestic abuse, coercive control, trauma, neurodiversity, allegations of harm, contested evidence, cultural context and serious consequences for a child’s safety and family relationships. No AI tool should be treated as a substitute for proper safeguarding analysis.

2. Am I sharing the right information?

This may be the most important question of all.

Before entering material into an AI system, consider:

  • Does the prompt contain a person’s full name?
  • Does it include a child’s identity or date of birth?
  • Does it contain a home address, school, medical provider or contact details?
  • Does it reveal domestic abuse, sexual allegations or health information?
  • Does it contain confidential client information?
  • Does it reproduce a solicitor’s advice or privileged communication?
  • Does it contain police, Cafcass, social services or medical records?
  • Does it reproduce documents filed in private family proceedings?
  • Does it include information about a third party who has not consented?

The fact that information is already stored electronically does not mean it is safe or lawful to transfer it into a separate AI system.

Users should also avoid assuming that deleting names is always sufficient. A person may remain identifiable from the combination of location, occupation, family structure, dates, allegations and case history.

Data minimisation must happen before submission

Where AI use is appropriate, provide only what is genuinely necessary.

That may mean:

  • replacing names with neutral labels such as “Mother”, “Father” and “Child A”;
  • removing addresses, telephone numbers and identifying references;
  • summarising the relevant issue instead of uploading an entire bundle;
  • excluding unrelated medical, sexual or financial information;
  • using an approved enterprise system rather than a personal consumer account;
  • checking retention, training and privacy settings before use.
“Would I be comfortable sending this information to an unknown external provider?” is a useful starting question. In legal work, however, comfort is not enough. The user must also consider confidentiality, privilege, data protection, court restrictions and professional duties.

3. Is the prompt clear?

Poor prompts produce poor outputs. More importantly, vague prompts can conceal poor reasoning.

Genevieve’s checklist proposes a useful formula:

Role
Task
Context
Constraints
Output

In legal work, each of those elements matters.

Role

What function is the system being asked to perform? Is it organising evidence, identifying inconsistencies, simplifying language or producing a first draft?

Simply telling an AI system to “act as a senior barrister” does not transform it into one. A role instruction may affect the structure and tone of an answer, but it does not create professional competence, accountability or legal authority.

Task

Define the actual job. “Help with my case” is too broad. “Create a chronological table from these dated events without adding facts or drawing conclusions” is clearer and safer.

Context

AI cannot reliably understand the context it has not been given. At the same time, users should not respond by dumping an entire life history, confidential file or court bundle into the system.

The discipline lies in providing sufficient relevant context without excessive disclosure.

Constraints

Appropriate constraints might include:

  • do not invent facts;
  • do not alter quoted wording;
  • distinguish evidence from allegation;
  • identify missing dates;
  • do not make findings of fact;
  • use neutral, child-focused language;
  • flag anything requiring legal verification;
  • state where the source material does not support a conclusion.

Output

Specify the form required: chronology, schedule, table, letter, neutral summary, list of issues or questions for professional advice.

A defined format makes it easier to review the result and identify whether the system has departed from its instructions.

4. Can the AI output be trusted?

Not without checking.

Generative AI can produce polished, fluent and authoritative-sounding text that is incomplete, misleading or simply wrong. Its tone may create an impression of certainty that the underlying material does not justify.

In legal contexts, common risks include:

  • invented case citations;
  • incorrect quotations from judgments;
  • outdated procedural rules;
  • confusion between different jurisdictions;
  • overstatement of legal tests;
  • failure to recognise exceptions;
  • miscalculated deadlines;
  • incorrect assumptions about the content of an order;
  • turning disputed allegations into apparent facts;
  • omitting evidence that does not fit the requested narrative.

Human review must be substantive. It is not enough to read an answer and think that it “sounds right”.

A proper verification process

  1. Check every legal proposition against a reliable current source.
  2. Open and read every cited judgment rather than trusting the summary.
  3. Verify all dates, figures, names and quotations.
  4. Compare the output against the original evidence.
  5. Check that allegations have not been presented as findings.
  6. Ask what relevant material may have been omitted.
  7. Ensure a responsible human approves the final document.

5. Could the prompt or output create risk?

Genevieve’s framework identifies four broad categories:

Security

Could the prompt contain malicious instructions, hidden content, unsafe links or prompt-injection material?

Privacy

Does the prompt involve personal, confidential, privileged or sensitive information?

Compliance

Could the use create bias, unfairness, unlawful processing or regulatory problems?

Governance

Must the prompt or output be recorded, reviewed, authorised or disclosed?

In litigation, a fifth category should be added: evidential and procedural risk.

Questions include:

  • Has the system changed the substance of a witness’s evidence?
  • Can the author explain and stand behind every sentence?
  • Has the output introduced facts that do not appear in the source material?
  • Has the use of AI affected authenticity or provenance?
  • Does the document comply with the relevant court rules, practice directions and orders?
  • Is disclosure of AI involvement required or appropriate?
  • Could the output mislead the court?

Prompt governance in family proceedings

Private family proceedings deserve particular attention because the underlying material is often intensely sensitive.

A typical case file may include:

  • children’s names, dates of birth, schools and medical information;
  • domestic abuse allegations;
  • sexual allegations;
  • police disclosure;
  • Cafcass safeguarding letters and section 7 reports;
  • social care records;
  • medical and therapeutic information;
  • private messages and photographs;
  • financial information;
  • information about third parties;
  • documents governed by reporting or publication restrictions.

Uploading an unredacted bundle to a general-purpose AI tool because it is convenient is not responsible case preparation.

This does not mean AI has no legitimate role. Used carefully, it may help litigants in person:

  • put events into chronological order;
  • identify repeated patterns of behaviour;
  • separate evidence from commentary;
  • improve the structure of a statement;
  • convert a long narrative into a schedule;
  • identify documents that appear to be missing;
  • prepare questions for legal advice or a hearing;
  • rewrite hostile correspondence into calm, child-focused language.

But the safeguards must come first.

Never ask AI to manufacture a stronger case

AI must not be used to embellish evidence, create allegations, invent conversations, alter screenshots, misrepresent legal advice or produce a false appearance of independent corroboration.

A witness statement must remain the witness’s truthful evidence. The person signing it must understand, approve and be able to defend its contents.

The professional position for legal services

Legal professionals are not prohibited from using artificial intelligence. But using a technological tool does not displace professional responsibility.

Solicitors and firms remain responsible for:

  • competence and service quality;
  • client confidentiality;
  • legal professional privilege;
  • data protection compliance;
  • accuracy of legal work;
  • supervision of staff and systems;
  • duties to the court;
  • acting in clients’ best interests;
  • ensuring that the court is not misled.

An organisation should therefore know which tools its staff are using, what information is being entered, what contractual and privacy terms apply, how outputs are checked and who remains accountable.

“A member of staff used ChatGPT” is not a governance framework.

A JSH Law “before you press submit” check

Before entering information

  1. Purpose: What exactly am I asking the system to do?
  2. Suitability: Is AI appropriate for this task?
  3. Authority: Am I permitted to use this tool and this information?
  4. Necessity: Does the system genuinely need all this material?
  5. Identity: Can names, addresses and identifying details be removed?
  6. Sensitivity: Does the material concern children, health, abuse, sexuality, criminal allegations or safeguarding?
  7. Confidentiality: Is any part confidential, privileged or restricted by the court?
  8. Security: Do I understand where the information will be processed and retained?

Before using the output

  1. Accuracy: Have all facts, calculations and legal propositions been checked?
  2. Evidence: Does every factual statement come from the source material?
  3. Neutrality: Have allegations and findings been clearly distinguished?
  4. Currency: Is the law and procedure up to date?
  5. Omissions: Has relevant contrary or qualifying material been left out?
  6. Responsibility: Can a named human stand behind the final document?
  7. Record: Should the prompt, output and review process be documented?
  8. Disclosure: Does the context require transparency about AI use?

Good governance should enable responsible use, not prevent it

Responsible AI governance is sometimes presented as an obstacle to innovation. That is the wrong way to look at it.

Clear rules allow people to use technology with greater confidence. They reduce uncertainty, protect sensitive information and make it easier to identify when human intervention is required.

The goal should not be to surround ordinary users with impenetrable policies. It should be to create practical safeguards that work at the point of use.

Genevieve Cripps’ checklist succeeds because it converts broad principles such as security, privacy, accuracy and accountability into questions a real person can ask before and after using AI.

That is where responsible adoption begins: not in a glossy strategy document, but in everyday decisions.

What this means for litigants in person

Litigants in person are already using generative AI. That reality cannot be wished away.

For someone who cannot afford extensive legal representation, AI may provide meaningful help with organisation, language and preparation. It may reduce the disadvantage caused by unfamiliar court processes and dense legal terminology.

But access to technology is not the same as access to reliable legal support.

Litigants in person should treat AI as a drafting and organisational assistant, not as an invisible lawyer, judge, safeguarding professional or source of unquestionable authority.

The safest approach is:

  • remove identifying and sensitive information wherever possible;
  • use AI for defined, limited tasks;
  • retain the original source documents;
  • check every substantive statement;
  • seek qualified advice where the issue is serious or complex;
  • never file material that you do not understand or cannot verify.

Conclusion

The prompt is not an inconsequential box of text. It can determine what data enters a system, what assumptions shape the result and what risks follow.

In legal and family justice settings, those risks are amplified because the information may affect rights, reputations, safety, children’s welfare and the fairness of court proceedings.

Prompt governance therefore needs to become part of basic professional and digital competence.

Before pressing submit, ask:

Should I use AI for this task?
Should I share this information?
Can I verify the result?
And am I prepared to remain accountable for what happens next?

Need help organising a legal case responsibly?

JSH Law provides practical, evidence-led support for litigants in person who need help turning large, disorganised or overwhelming case material into clear documents for use in family proceedings.

Support may include:

  • chronologies and schedules of events;
  • witness statement structure and review;
  • evidence organisation;
  • Cafcass report analysis;
  • hearing preparation;
  • appeal paperwork;
  • non-molestation order applications;
  • identifying gaps, inconsistencies and safeguarding issues;
  • responsible use of AI-assisted legal preparation.

The purpose is not to manufacture a case. It is to present the evidence accurately, calmly and effectively, while keeping the child’s welfare and the court’s decision-making needs firmly in view.

Book a JSH Law consultation

Sources and further reading

Disclaimer: This article provides general legal and practical information for England and Wales. It does not constitute legal advice and should not be relied upon as a substitute for advice about the facts of an individual case. The legal, procedural, confidentiality and data protection implications of using AI depend on the tool, information, purpose and circumstances involved.

© JSH Law. This article includes independent commentary on the work of Genevieve Cripps. Genevieve’s original checklist remains her own work and should be credited and linked whenever referenced.

The JSH Law Legal Tech Test: Which Platforms Actually Improve Access to Justice?

/0 Comments/in , , /by

Legal technology is not automatically access to justice. A platform may look simple, generate polished documents and promise to reduce legal costs—but the real test is whether it helps an ordinary person protect their position without being misled, exposed or given false confidence. The JSH Law Legal Tech Test will examine that question properly.

JSH Law Legal Technology Review Series

The JSH Law Legal Tech Test: Which Platforms Actually Improve Access to Justice?

Published 3 August 2026  |  Research current at the date of publication

Legal technology should not be judged by how impressive its artificial intelligence sounds. It should be judged by whether a real person can use it to understand the process, protect their position and reach the right next step without being misled, exposed or priced out.

Legal technology is spreading rapidly across the justice system. Some platforms draft court documents. Some organise evidence, track deadlines or manage communication between separated parents. Others offer legal research and litigation analysis that remains largely available only to firms, barristers and well-funded clients.

That growth matters. Used well, technology can reduce cost, delay and procedural confusion. Used badly, it can give a frightened or inexperienced person false confidence in a document, deadline or legal route that has never been properly checked.

JSH Law is therefore launching a continuing review series: The JSH Law Legal Tech Test. We will examine the websites, apps, court services and AI platforms that claim to make law easier, cheaper or more accessible. The first full review will be CaseCraft AI, a platform aimed at people bringing or defending small claims in England and Wales.

Legal technology is not automatically access to justice

A clear questionnaire can be genuinely valuable. It can prompt a user to identify dates, parties, sums, documents and missing evidence. Automated reminders can prevent a deadline from disappearing inside an inbox. A well-designed platform can turn an intimidating process into a sequence of manageable tasks.

But a smoother interface does not remove legal complexity. It can simply conceal it.

The small claims track, for example, is normally used for many straightforward claims worth no more than £10,000. That is not the whole rule. Different thresholds and exceptions apply to some personal injury cases; certain claims are not normally allocated to the track; and allocation remains a case-management decision for the court. The official framework is found in CPR Part 26 and its accompanying practice direction.

A platform can generate a polished-looking claim form and still fail the user if it does not identify limitation, jurisdiction, the correct defendant, a mandatory pre-action step, an unsuitable cause of action, a counterclaim or an enforcement problem. In law, presentation is not reliability.

The central question

Does the technology reduce the user’s legal and procedural risk—or merely make the process feel easier?

We will also distinguish between a regulated law firm, an unregulated technology provider, a charity, a government court service and a general-purpose AI tool. They do not carry the same duties, redress arrangements or consumer protections. A familiar design and confident language must never be mistaken for regulation.

Why CaseCraft AI will be reviewed first

CaseCraft is an unusually useful starting point because it is not simply offering a chatbot or a document template. Its website presents an end-to-end small-claims workflow for claimants and defendants: guided information gathering, document generation, evidence organisation, filing, deadline tracking and support with settlement and enforcement.

Its public-facing proposition is attractive. It says a user can begin without paying a traditional hourly legal fee and that a success-based commission applies to a favourable outcome. Its website also identifies Sterling Lawyers Ltd, SRA number 630147, as the regulated firm providing legal services through the platform.

That combination—AI automation, court documents, a success fee and a regulated legal-services relationship—raises exactly the questions this series is designed to investigate.

What the full CaseCraft review will test

  • which claims and defences the platform accepts, rejects or refers for human review;
  • whether its eligibility questions identify limitation, jurisdiction, track-allocation and pre-action issues;
  • what documents it produces and whether they are factually accurate, properly pleaded and usable;
  • how counterclaims, multiple parties, disputed facts and cases that become more complex are handled;
  • which company contracts with the user and when a solicitor-client relationship begins;
  • the total price, including onboarding charges, commission, court fees, hearing fees and enforcement fees;
  • whether a commission becomes payable on an award that is never recovered;
  • what human checking, complaints process, insurance and Legal Ombudsman route apply;
  • how uploaded evidence and personal data are stored, used and deleted; and
  • whether a vulnerable, disabled or digitally excluded user can obtain meaningful human assistance.

An early reason for careful scrutiny

Before conducting a hands-on test, a desk review of CaseCraft’s public pages already shows why legal-tech terms must be read rather than inferred from a headline.

The main website currently advertises a 15% success fee. CaseCraft’s terms, stated to have been updated on 27 July 2026, also refer to a £15 onboarding fee, court fees paid in advance and a 15% commission in specified claimant and defendant outcomes. The terms say that, in some claimant cases, commission may remain payable even if an award cannot be enforced. However, other public CaseCraft promotional or editorial pages visible at the date of review have referred to a 10% fee.

There is also wording that deserves clarification about the contracting and regulatory structure. The terms describe the agreement as being with Sterling Lawyers Ltd trading as CaseCraft, while the website footer describes CaseCraft AI Ltd as the platform operator and Sterling Lawyers Ltd as a separate, independent affiliate providing legal services.

This is not a finding that the service is unsafe or that any particular fee will be charged. The terms presented during the actual sign-up journey, the scope accepted by the regulated firm and the individual client documentation will matter. It is, however, a clear reason to test price transparency and regulatory accountability carefully in the full review.

JSH Law will invite CaseCraft to clarify those points and, if possible, provide a demonstration or review account. Any substantive response will be included fairly. The final article will state whether it is based on a hands-on test, a guided demonstration or public information only.

The JSH Law 100-point legal-tech test

Every reviewed platform will be assessed against the same core standard. A product will not receive a high score merely because it is fast, attractive or powered by a sophisticated model. The scoring gives the greatest weight to reliability, accountability and the treatment of sensitive information.

Test Weight What we will examine
Legal and procedural reliability20Accuracy, jurisdiction, deadlines, legal tests, forms, authorities, warnings and escalation of uncertainty.
Practical usefulness15Whether the product helps a user complete the real task, not merely generate text.
Regulation and accountability15Provider identity, regulated status, scope, insurance, complaints, redress and responsibility for errors.
Privacy and data security15Data controller, hosting, retention, deletion, model training, third-party access and treatment of sensitive evidence.
Accessibility and ease of use10Plain English, disability access, digital confidence, mobile use and clarity when something goes wrong.
Price and value10Total cost, additional fees, renewal, cancellation, recoverability and value compared with free or human alternatives.
Human assistance10When human review is available, who provides it, their qualifications and whether urgent escalation works.
Safeguarding and trauma awareness5Recognition of abuse, coercion, vulnerability, unsafe joint working, litigation misuse and risk to children.
Total100A published score supported by reasons, limitations and evidence.

Three evidence labels

Every article will carry one of the following labels so readers know what has—and has not—been independently verified:

Hands-on test
JSH Law used the service through a structured test journey.
Guided demonstration
The provider demonstrated the platform and answered questions.
Desk-based assessment
The review relies on public pages, terms, policies, official records and other identified sources.

Where a live service requires a real legal problem, payment, identity verification or the upload of personal evidence, we will not pretend to have completed a transaction that did not occur. Any test case will use fictional or properly anonymised material.

The first 15 platforms in the JSH Law review queue

The opening series will compare commercial products with regulated services, court systems, charities and public legal-information projects. That comparison matters. Sometimes the best access-to-justice technology is not the product with the largest AI claim. It is the service that knows its limits and gets the user safely to the next step.

Order Platform The question JSH Law will answer
1CaseCraft AICan an AI-led platform make bringing or defending a small claim genuinely safer and simpler—and are its price, regulation and limits clear?
2GarfieldWhat difference does an SRA-regulated AI debt-recovery model make to accountability and user protection?
3CourtNavCan a free guided service help a domestic-abuse survivor prepare an injunction application without losing the safety of human legal review?
4VallaDoes combining case-management tools, templates and pay-as-you-go human coaching create a workable model for self-represented tribunal users?
5amicableWhen is one service for a separating couple efficient, and when do conflict, imbalance or domestic abuse require separate advice?
6OurFamilyWizardCan recorded co-parenting communication reduce conflict, or can a high-conflict user turn the app into another channel of pressure and surveillance?
7AdvicenowWhat does effective digital help for litigants in person look like when it is designed around explanation rather than AI marketing?
8LawhiveDoes an AI-enabled consumer law-firm model widen affordable access to a lawyer while preserving quality and individual judgment?
9ResolverCan guided complaints and record-keeping resolve consumer disputes before court becomes necessary?
10Support Through CourtWhich parts of court support still depend on a calm, trained human being listening to the person behind the paperwork?
11Online Civil Money ClaimsIs the government’s online money-claim route genuinely designed around the needs and limitations of litigants in person?
12Online divorceDoes a simpler divorce application risk users assuming that children, finances, housing and safeguarding have also been resolved?
13Legal Aid CheckerDoes the digital eligibility journey help vulnerable applicants find a provider, or does it merely tell them that help might exist?
14Rocket LawyerWhat does a consumer actually receive from a legal-document subscription, AI assistance and access to a lawyer?
15LawDepotWhen is a self-generated legal document useful, and when does a template hide the need for advice or bespoke drafting?

Family justice and co-parenting technology need a different test

Family technology cannot be assessed as though every case involves two safe, equally powerful adults who simply need a better shared calendar.

In a cooperative separation, a co-parenting app may centralise dates, expenses, messages and child-related information. In a case involving coercive control, stalking, harassment or litigation abuse, the same functions can have a very different effect. Read receipts can become a demand for immediate compliance. Location tools can create fear. An immutable record can protect one parent—or supply the other with a new arena for performative, controlling communication.

The test is not whether an app encourages a polite tone. It is whether its design understands power, safety and the difference between ordinary disagreement and abuse.

If AI can guide someone through a small money claim, the family-justice question is unavoidable: why are parents still expected to organise years of safeguarding evidence, identify patterns of coercive behaviour and comply with complex directions with so little structured support?

The family and separation watchlist will therefore include amicable co-parenting, 2houses, TalkingParents, AppClose, Settify and Class Legal’s Capitalise.

Where a product is principally built or marketed for another jurisdiction, we will say so. A claim that an app is used in American courts does not establish its evidential status, procedural suitability or judicial treatment in England and Wales.

Evidence, bundles and professional legal AI

The next group exposes a growing inequality of arms. Professional teams increasingly have access to tools that can search large document sets, build chronologies, identify contradictions, create bundles and accelerate legal research. An unrepresented person may still be copying messages into a spreadsheet at two o’clock in the morning.

Our evidence and litigation reviews will consider platforms including TrialView, Opus 2, Legora, Thomson Reuters Case Center, Bundledocs and Casedo.

Professional legal-AI reviews will include Lexis+ with Protégé, CoCounsel Legal UK, Vincent by vLex, Harvey, Luminance, Genie AI, Spellbook, Clio and LEAP.

The purpose is not to demand that an enterprise product be sold to every litigant. It is to ask a policy question: if technology can make complex evidence understandable for a commercial team, which parts of that capability could be made safe and affordable for people navigating child arrangements, domestic abuse, housing or employment proceedings alone?

The general AI tools people are already using

We will also test ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot and NotebookLM. These are not substitutes for a lawyer, but people are already using them to understand orders, draft statements, summarise evidence and ask legal questions.

They will be tested against the same fictional case materials and controlled questions. We will examine fabricated authorities, missing jurisdictional caveats, false certainty, source quality, confidentiality warnings, handling of contradictory evidence and whether the system recognises when a safeguarding issue needs human help.

Before using any legal-tech platform: ten questions to ask

  1. Who is the contracting party? Record the company name, address and registration details.
  2. Is anyone providing a regulated legal service? Check the regulator’s register yourself and identify the exact firm.
  3. What is included? “Guidance”, “document preparation”, “legal advice” and “representation” are not the same service.
  4. What is the full cost? Include subscription, onboarding, percentage fees, court fees, hearing fees, expert fees and enforcement.
  5. When does payment become due? A favourable order is not necessarily money recovered.
  6. Who checks the output? Ask whether review is automated, administrative, legally qualified or supervised by a regulated professional.
  7. What happens when the case stops being simple? Look for a clear escalation, referral and exit process.
  8. What happens to the evidence? Check storage, retention, deletion, overseas transfers, model training and third-party processors.
  9. What remedy exists if it goes wrong? Find the complaints route, ombudsman, insurance position and governing law.
  10. Is it safe for this particular case? Consider domestic abuse, coercive control, child data, confidentiality, capacity, disability and digital exclusion.

Legal files may contain health information, allegations of abuse, sexual information, children’s details and criminal-offence data. The Information Commissioner’s Office guidance on AI and data protection emphasises lawfulness, transparency, fairness, accuracy, security, data minimisation and accountability. “Secure” is not a complete privacy explanation.

What genuine access-to-justice technology should look like

The best legal technology will not try to erase professional judgment. It will use technology for what technology does well: structure, prompts, consistency, search, reminders, comparison and organisation. It will use trained human beings for what still requires judgment: disputed facts, legal strategy, risk, credibility, vulnerability, proportionality and safeguarding.

It will be clear about who is responsible. It will not hide a material fee in a long document or describe an uncertain outcome as inevitable. It will not assume that every user is confident, safe, literate, represented or able to pay. It will make the court’s job easier by helping the user present the relevant facts and evidence—not by producing more polished noise.

That is the standard this series will apply.

Do you use a legal-tech platform?

JSH Law wants to hear from litigants in person, practitioners, charities and platform providers. Tell us what worked, what failed, what the service cost and what you wish you had known before relying on it. Providers are welcome to offer a demonstration and respond to the review questions.

Contact JSH Law

Research and source note

This launch article is a desk-based assessment and editorial roadmap, not a completed product review or endorsement. Public websites and terms can change. Pricing and contractual terms should be checked immediately before purchase.

JSH Law has not received payment for including the platforms in this article and the order is editorial, not a ranking or recommendation. Any future commercial relationship, complimentary access or provider-assisted demonstration relevant to a review will be disclosed.

Important: This article provides general legal information and commentary for England and Wales. It is not legal advice and does not create a solicitor-client relationship. The correct procedure and suitability of any service depend on the facts, the documents, the court or tribunal, the terms in force and the remedy sought. Obtain appropriate legal advice where you are unsure, where a deadline is approaching, or where the case involves significant loss, domestic abuse, safeguarding, children, capacity or complex evidence.

Family Law Technologists: Where AI Meets Evidence, Safeguarding and Justice. My Two Penneth.

/0 Comments/in , , /by

Artificial intelligence is already changing how legal work is researched, organised and presented. For litigants in person, it may offer an affordable way to understand procedure, prepare chronologies and improve court documents. But in family law—where decisions may affect children, safety, homes and family relationships—an answer that merely sounds convincing is not enough.

Artificial Intelligence, Family Law and Access to Justice

The Fine Structure of Justice: What Physics Can Teach Us About AI, Family Law and Litigants in Person

Artificial intelligence is already helping to prepare court cases, organise evidence, support judges and guide people who cannot afford conventional legal representation. But legal AI needs governing constants of its own: accuracy, accountability, confidentiality, transparency, human oversight and access to redress.

By Jessica Susan Hill | JSH Law | Published 24 July 2026

Gold alpha symbol and fine-structure constant equation surrounded by luminous atomic geometry on a dark indigo background
The fine-structure constant describes a fundamental interaction in physics. Legal AI now needs dependable principles governing the interaction between technology and justice.

The short answer

AI can make legal help cheaper, faster and easier to understand. It can be particularly valuable to litigants in person who need help organising evidence, preparing chronologies and understanding court procedure. But it cannot safely replace verified legal research, professional responsibility, safeguarding judgment, human advocacy or judicial independence. The future should be AI-supported, human-accountable and evidence-led justice.

For several years, the legal profession discussed artificial intelligence as though it were standing outside the courtroom, waiting to be admitted.

That conversation is now out of date.

AI is already inside law firms, government departments and the justice system. It is being used to research legal questions, review documents, produce transcripts, organise evidence and support judicial administration. Members of the public are using ChatGPT, Claude, Gemini and Copilot to understand their legal problems before they ever speak to a lawyer—if they speak to one at all.

In May 2026, the Master of the Rolls said that AI would be used in every aspect of the work of lawyers and judges. He predicted that it would allow legal and judicial work to be performed more quickly and at more proportionate cost.

The question is therefore no longer whether AI will be used in law. It is what kind of legal system we are building around it.

What does the fine-structure constant have to do with justice?

In physics, the fine-structure constant is represented by the Greek letter α, or alpha. It is a dimensionless number, approximately equal to 1/137, which measures the strength of electromagnetic interaction between charged particles.

It helps physicists describe how matter interacts with light and how the fine structure of atomic energy levels arises. It is a small number with enormous consequences.

Legal artificial intelligence needs governing constants of its own.

I do not mean one mysterious numerical value. I mean a set of principles that must remain fixed whenever technology interacts with legal rights, evidence and human vulnerability:

The six constants of responsible legal AI

  1. Accuracy: legal propositions, authorities, dates and procedural requirements must be capable of verification.
  2. Accountability: an identifiable human being or regulated organisation must remain responsible for consequential legal work.
  3. Confidentiality: court papers, children’s information, medical records and privileged communications must be protected.
  4. Transparency: users must understand when AI has been used and what its limitations are.
  5. Human oversight: technology must support rather than displace legal judgment, safeguarding assessment and judicial independence.
  6. Access to redress: people harmed by defective legal AI need somewhere effective to complain and obtain a remedy.

These principles matter in every legal field. They become especially important where somebody is representing themselves.

A litigant in person may ask a general-purpose chatbot to interpret a court order, identify the relevant law or prepare a witness statement without knowing whether the response is accurate. In family proceedings, the information may concern domestic abuse, coercive control, children’s welfare, confidential medical evidence or an urgent safeguarding risk.

The challenge is therefore not simply to make legal AI more powerful. It is to calibrate the relationship between machine assistance and human justice.

Get that relationship right and AI could widen access to justice. Get it wrong and we may automate error, inequality and harm at unprecedented speed.

AI has already entered the English courtroom

In May 2026, a freelance HR consultant reportedly recovered approximately £7,000 in unpaid fees following a three-hour trial at Wandsworth County Court. She had used Garfield AI, an SRA-authorised AI-centred law firm, for the pre-trial legal work.

The reported cost of that assistance was approximately £400. Garfield AI helped with the court documents, witness statements and trial bundle. A human barrister conducted the advocacy.

That distinction matters.

This was not an autonomous machine standing before a judge and winning a case. It was an example of technology performing repeatable preparation work while a human advocate remained responsible for presenting and testing the case.

Nevertheless, it demonstrated something important. Claims that were previously uneconomic to pursue may become viable if the cost of legal preparation falls.

For a freelancer owed £7,000, a tenant in dispute with a landlord, an employee pursuing unpaid wages or a parent needing help to organise a family-court application, that change is not theoretical. It may determine whether they can enforce their rights at all.

This should not be treated as proof that lawyers are obsolete. Nor should it be dismissed as a publicity exercise. It is evidence that the economics of legal work are changing.

Further reading: report on the Garfield AI-assisted county court case .

The justice system itself is adopting AI

In June 2026, the Ministry of Justice announced plans to develop and test AI legal assistants to support routine casework, legal research and case analysis. A further tool is intended to help judges identify trial-ready cases and group similar hearings, with the stated aim of reducing delays in the Crown Court.

The government says the technology will first be tested in controlled environments against standards for safe and ethical use.

That is essential. The justice system handles criminal allegations, children’s information, domestic-abuse evidence, medical records and material capable of changing the course of a person’s life.

AI is also being used in more targeted judicial work. In April 2026, the Chancellor of the High Court described how AI was helping judges identify information that may need to be removed from published judgments to prevent individuals from being identified.

This is particularly relevant to family cases. Removing names may not be enough. A combination of locations, relationships, medical conditions, school information or unusual factual details can still identify a child or family. AI may help detect that “jigsaw identification” risk.

The final responsibility, however, remains with the judge.

What AI must not become

AI must not become a convenient technological answer to chronic underfunding, insufficient court staff, legal-aid deserts or overwhelming backlogs. It may help professionals work more effectively, but it cannot manufacture judicial capacity, procedural fairness or public trust.

Sources: Ministry of Justice, AI tech ambition to deliver smarter justice for victims; Chancellor of the High Court, Legal Professional Privilege in the Age of AI; and Master of the Rolls, Artificial Intelligence and the Judiciary.

AI could transform access to justice—but the protection gap is real

The strongest argument for legal AI is not that it will make already profitable law firms even more profitable. It is that it may provide meaningful assistance to people who currently receive none.

Research published by the Legal Services Board in June 2026 found that consumers were broadly optimistic about the possibilities:

  • 70% expected AI to make legal services easier to use;
  • 66% expected improved accessibility; and
  • 64% expected legal services to become more affordable.

That optimism came with conditions. Consumers expected accuracy, human oversight, informed consent, protection of personal information and a route to complain or obtain redress.

Those are not unreasonable demands. They are the minimum foundations of trustworthy legal assistance.

A client using a regulated legal service ordinarily benefits from professional duties, supervision, insurance and a complaints process. A person relying directly on a general-purpose chatbot may have none of those protections. They may not even realise that there is a difference.

AI should reduce the cost of obtaining reliable help. It should not simply transfer the cost of mistakes to the person least able to bear them.

Source: Legal Services Board, AI tools show real promise to increase access to legal services .

AI hallucinations are not a minor technical inconvenience

Generative AI can produce information that sounds authoritative but is inaccurate, incomplete or entirely invented. In law, this can include fictitious cases, fabricated quotations, incorrect legislation and false descriptions of what a judgment decided.

In Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank QPSC [2025] EWHC 1383 (Admin), the Divisional Court addressed false legal authorities placed before the court.

The judgment emphasised that legal representatives remain responsible for material submitted to the court. It also warned that public AI tools may invent cases, citations or quotations and may provide incorrect or misleading information about the law.

If you put it before a court, you remain responsible for it.

A confident answer is not necessarily a correct answer. A citation is not verified merely because it looks properly formatted. Every important legal proposition should be checked against the actual legislation, judgment, procedural rule or authoritative guidance.

That responsibility applies to lawyers. It also matters to litigants in person.

However, it is unrealistic to pretend that an unrepresented, distressed court user has the same research skills, database access or professional training as a regulated practitioner. Education, safer tools and affordable human support are more useful than simply warning people that they use AI at their own risk.

Source: Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank QPSC [2025] EWHC 1383 (Admin) .

Should AI-generated court documents be disclosed?

The Civil Justice Council has been examining whether additional rules are needed when legal representatives use AI to prepare pleadings, witness statements, expert reports and other court documents.

Its June 2026 update recorded substantial agreement that existing professional responsibilities may be sufficient for pleadings, skeleton arguments and similar documents. Witness statements remain a more difficult area.

That is because a witness statement is not simply a piece of legal writing. It is supposed to contain the witness’s own evidence and personal recollection.

AI may help a witness organise dates, remove repetition and express their account more clearly. But it may also reshape language, strengthen uncertain recollections, introduce details or turn a complicated human memory into an artificially seamless narrative.

That creates a particular risk when the underlying evidence is contested.

Assistance is not authorship

A litigant should not be criticised merely because technology helped turn disorganised information into a readable chronology or properly structured statement. The real questions are whether the evidence remains authentic, whether the witness understands and adopts it, and whether every factual assertion is accurate.

Disclosure rules should protect the integrity of evidence without stigmatising people who use accessible technology to overcome disability, trauma, language difficulties or lack of legal representation.

Source: Civil Justice Council, Use of AI in preparing court documents .

Why family law requires particular care

Family proceedings are not document-production exercises. They frequently involve domestic abuse, coercive control, disputed allegations, trauma, vulnerable adults, children’s wishes and feelings, safeguarding risk and intensely personal evidence.

AI can help a parent:

  • identify relevant dates;
  • organise messages and other evidence;
  • prepare a clear chronology;
  • remove repetition from a draft statement;
  • understand the basic structure of a position statement;
  • identify questions to ask Cafcass or a legal adviser; and
  • turn a large quantity of information into manageable categories.

Those functions can make an enormous difference to an overwhelmed litigant in person. They can also make the court’s job easier by helping the litigant present the real issues clearly and proportionately.

But AI cannot safely decide:

  • whether a child is at immediate risk;
  • whether apparently isolated incidents form part of a coercive pattern;
  • whether apparently reasonable language disguises controlling behaviour;
  • which allegations are relevant to welfare and future risk;
  • whether communication or direct contact with another party is safe;
  • what evidence should be disclosed, redacted or urgently preserved;
  • whether an agreement has been reached freely; or
  • what order is necessary and proportionate in a particular case.

General-purpose AI often produces polished, balanced and conciliatory language. That may be helpful in ordinary disputes. In domestic-abuse cases, however, it can flatten the reality of coercive control.

A sustained pattern of intimidation may be rewritten as a “communication difficulty”. Repeated breaches of boundaries may become a “difference in parenting styles”. A safety-driven refusal may be presented as an unwillingness to compromise.

That is not neutrality. It is the loss of legally and safeguarding-relevant context.

The opposite problem can also occur. AI may overstate weak evidence, apply legal labels too readily or encourage a parent to include every grievance they have ever experienced. That does not strengthen a case. It can bury the central welfare issues beneath unnecessary material.

Family-law AI must be trauma-informed and safeguarding-aware. It must recognise when a question has moved beyond document assistance and requires specialist human judgment.

How litigants in person can use AI more safely

AI can be genuinely useful to somebody representing themselves, provided it is treated as an assistant rather than an authority.

Good uses of AI

  • organising your own account into a dated chronology;
  • summarising correspondence you have already checked;
  • identifying repetition or unclear passages;
  • turning a long draft into structured headings;
  • preparing a list of questions for a legal adviser;
  • explaining unfamiliar terminology in plain English;
  • creating a document or hearing checklist; and
  • improving the readability of words you have written.

Matters requiring independent verification or human advice

  • case citations and quotations from judgments;
  • filing dates and limitation periods;
  • the precise legal effect of an existing order;
  • the admissibility or relevance of evidence;
  • the correct application, form or procedural route;
  • the likelihood of a particular outcome;
  • immediate safeguarding or domestic-abuse risk; and
  • complex, urgent or contested proceedings.

A practical five-step verification rule

  1. Ask the AI to identify the source for every important legal proposition.
  2. Open and read the source yourself.
  3. Confirm that the source applies in England and Wales and remains current.
  4. Check every factual statement against your own evidence.
  5. Obtain human advice where an error could affect a child, your safety, your liberty, your home or a significant sum of money.

Legislation should be checked on legislation.gov.uk. Court rules and official guidance should be checked through GOV.UK or the Judiciary website. A case name should be verified by locating and reading the actual judgment.

Never cite a case merely because an AI system supplied its name.

AI will change the legal business model as much as the legal work

The most disruptive feature of AI may not be its ability to draft. It may be its effect on time.

If work that once took five hours can be completed responsibly in one, clients will increasingly question why they should be charged according to the old timescale.

Law firms will need to demonstrate value through judgment, strategy, accountability, advocacy and outcomes—not simply through the volume of time recorded.

That does not make lawyers less important. It changes where their value lies.

The most valuable legal professionals will be those who can use AI efficiently while recognising where it is unreliable; who can distinguish information from evidence; and who remain prepared to take responsibility for the work placed before a client, opponent or court.

The more autonomous the technology becomes, the clearer human responsibility must become.

The fine structure of responsible legal AI

The legal profession does not need to choose between blind enthusiasm and defensive resistance. It needs enforceable standards, responsible innovation and a clear understanding of what must remain human.

AI should make legal expertise more accessible. It should not disguise the withdrawal of that expertise.

It should reduce unnecessary cost. It should not remove accountability.

It should help judges and lawyers understand evidence. It should not decide whose evidence is believed.

It should help litigants in person communicate clearly. It should not rewrite their memories or invent their legal authorities.

It should identify patterns. It should not reduce domestic abuse, coercive control or safeguarding risk to decontextualised data points.

The fine-structure constant helps describe the interaction between matter and light. The fine structure of justice will depend on how carefully we govern the interaction between artificial intelligence and human judgment.

The governing constants must be accuracy, accountability, confidentiality, transparency, human oversight and redress.

If those principles remain fixed, AI may become one of the most important access-to-justice tools of our generation.

If they are treated as optional, AI will not correct the inequalities already present in our legal system. It will reproduce them faster.

Frequently asked questions about AI and law

What is the fine-structure constant?

The fine-structure constant, represented by α, is a dimensionless physical constant measuring the strength of electromagnetic interaction. Its value is approximately 1/137. This article uses it as a metaphor for the fixed principles needed to govern the interaction between AI and justice.

Is AI already being used by UK courts?

Yes. Secure AI tools are being used or tested for particular judicial and administrative tasks, including anonymisation, research, analysis and case management. Judges remain responsible for their decisions and judgments.

Can ChatGPT help a litigant in person?

It can help organise information, prepare chronologies, explain terminology and improve the structure of a draft. It should not be treated as an authoritative source of law, relied on for safeguarding decisions or used without checking its work.

Can AI prepare a family-court witness statement?

AI can assist with structure and clarity, but the statement must remain the witness’s own evidence. Every factual assertion must be checked, the witness must understand and adopt the statement, and AI must not embellish or reconstruct their recollection.

Must lawyers disclose that they have used AI?

There is no universal requirement to disclose every administrative or drafting use. The Civil Justice Council is continuing to consider whether further safeguards are required, particularly where AI has been used in preparing witness statements or evidence.

Is it safe to upload court documents to a public AI chatbot?

Not automatically. Court documents may contain confidential, personal or highly sensitive information. Users must understand the provider’s storage, processing, training and deletion arrangements and should remove identifying information where appropriate.

Will AI replace solicitors, barristers or judges?

AI is likely to reduce the time spent on repetitive research, review and document-production tasks. It is far less suited to professional responsibility, advocacy, negotiation, credibility assessment, safeguarding and judicial judgment.

JSH Law: using technology without losing sight of the human case

AI can help organise information, but effective court preparation still requires judgment: identifying the real issues, checking the evidence, removing material that weakens the case and presenting the position clearly and proportionately.

JSH Law provides practical, evidence-led and safeguarding-aware support with chronologies, witness statements, schedules of allegations, position statements, Cafcass material, appeal paperwork, court bundles and hearing preparation.

If you are representing yourself, the aim is not to make your case sound more legal. It is to make the relevant facts, evidence, welfare concerns and orders sought easier for the court to understand.

Book a consultation

About the author

Jessica Susan Hill writes about artificial intelligence, legal technology, family justice, domestic abuse, safeguarding, litigants in person and access to justice. Her work focuses on how technology can make legal processes more understandable and accessible without weakening professional responsibility, evidence quality or human judgment.

Legal information notice: This article provides general information about artificial intelligence and legal services in England and Wales as at 24 July 2026. It is not legal advice. Legal procedure and the appropriate use of technology depend on the facts, the jurisdiction, applicable court rules and the nature of the information involved.

AI Has Helped Win a UK Court Case — But It Is Not the Story Some Headlines Suggest

/0 Comments/in , , /by

The headlines about an AI-powered legal win in the UK are eye-catching, but they need to be read carefully. Garfield AI’s reported success is a genuine legal technology milestone, but it is not the moment AI replaced lawyers in court. The more important point is that regulated AI may now be helping ordinary people pursue legal claims that would otherwise be too expensive, stressful or time-consuming to bring.

AI, Legal Tech and Access to Justice

AI Has Helped Win a UK Court Case. But Let’s Be Clear What That Really Means.

Garfield AI’s reported court success is a genuine legal technology milestone. But it is not quite the “AI lawyer replaces humans” story some headlines suggest. The real significance is more practical, and potentially more important: AI may be starting to make low-value legal claims economically viable again.
Updated: June 2026
Focus: Garfield AI, SRA regulation, small claims, access to justice, legal AI, family law, litigants in person and the future of legal services in England and Wales.

A recent report about an AI-powered legal win in the UK has been circulating widely. The headlines are eye-catching. An AI law firm. A successful court case. A first for England and Wales. A legal technology milestone.

The story matters. But it also needs to be understood properly.

The important point is not that artificial intelligence walked into court and replaced a barrister. It did not. The reported case involved Garfield AI preparing the pre-trial legal work in an unpaid debt claim of around £7,000, with a human barrister conducting the advocacy at Wandsworth County Court.

That distinction matters.

This is not a binding precedent from the Court of Appeal or Supreme Court. It does not change the law. It does not mean AI can represent people in court on its own. It does not mean lawyers are suddenly redundant.

But it is still significant.

The real story is not “AI replaces lawyers”. The real story is that regulated AI may help ordinary people and small businesses pursue claims that would otherwise be too expensive, too stressful or too time-consuming to bring.

What actually happened?

Public reporting describes Garfield AI, an SRA-authorised AI-driven law firm, assisting a freelance HR consultant in recovering unpaid fees of around £7,000. The claim was heard at Wandsworth County Court, and the AI system is said to have prepared the legal documents and pre-trial material. A human barrister then conducted the court advocacy.

According to Garfield AI’s own statement, the case involved a freelancer defeating a counterclaim and recovering the unpaid sum. The Guardian also reported that the client paid Garfield AI around £400 for the process, which is precisely why the story has attracted so much attention.

This is the kind of case that often falls into the access to justice gap. The sum is large enough to matter deeply to the person owed the money, but not always large enough to justify traditional legal fees.

That is the problem legal AI is trying to solve.

Read the original reporting and source material:

What this story does not mean

Before anyone gets carried away, we need to be precise.

This is not a binding legal precedent.

Some headlines have described the case as setting a landmark precedent. In ordinary media language, it may be a landmark moment. But in legal terms, a County Court small debt claim does not create binding authority for other courts.

That does not make it unimportant. It simply means we should not overstate it.

The case is better understood as a legal services milestone, not a doctrinal legal precedent.

It does not mean:

  • AI appeared in court on its own.
  • AI replaced the judge.
  • AI replaced advocacy.
  • AI created new law.
  • AI can safely handle every type of claim.
  • AI can be used without regulation, supervision or safeguards.

What it does mean is more interesting.

It shows that AI-assisted legal preparation, within a regulated structure, may be capable of supporting low-value litigation that many people would otherwise abandon.

Why the SRA authorisation matters

Garfield AI is not just a random chatbot operating outside the legal system. The Solicitors Regulation Authority authorised Garfield.Law Ltd in 2025 as the first purely AI-based firm providing regulated legal services in England and Wales.

That matters because regulation is central to the legal AI debate.

There is a huge difference between:

  • a person asking a public chatbot for help with a court form;
  • a non-regulated document tool generating legal-looking text;
  • a solicitor using AI privately without proper checking;
  • and an SRA-authorised AI-based legal service operating under regulatory duties and safeguards.

The SRA’s approval of Garfield AI was not a blank cheque for AI. It was significant because it brought the tool inside the regulated legal services framework.

The key point: AI in legal services cannot be judged only by whether the output looks impressive. The real questions are: who is responsible, what safeguards exist, how is accuracy checked, how is client data protected, and what happens when something goes wrong?

Why this matters for access to justice

The access to justice point is the most important part of this story.

Many people and small businesses do not pursue valid claims because the cost, time, stress and uncertainty of litigation outweigh the amount at stake.

That is true in debt claims. It is true in consumer disputes. It is true in housing problems. It is true in employment issues. And, in a different way, it is also true in family law.

The justice system contains a vast number of people who need legal help but cannot afford traditional full-service representation.

They are not necessarily looking for a magic robot lawyer. Often, they need something much more practical:

  • help understanding the process;
  • help organising the facts;
  • help drafting clear documents;
  • help knowing what evidence matters;
  • help meeting deadlines;
  • help preparing for a hearing;
  • help avoiding procedural mistakes;
  • help deciding whether a claim is worth bringing at all.

If AI can reduce the cost of that support, it could make a real difference.

The access to justice opportunity:

AI may help make legal support available for cases that are currently uneconomic for traditional solicitors and too complex for ordinary people to handle comfortably alone.

The family law angle

Although the Garfield AI case was not a family law case, family lawyers should still pay attention.

Legal technology usually reaches family law later than commercial law or debt recovery, but it does reach us eventually.

The family justice system already has thousands of litigants in person. Many cannot afford solicitors. Many cannot get legal aid. Many are trying to manage emotionally overwhelming proceedings while also preparing documents, evidence, chronologies, statements and court forms.

That means the demand for AI-assisted legal support in family cases will grow.

But family law is not a simple debt claim.

Family cases may involve:

  • children;
  • safeguarding;
  • domestic abuse;
  • coercive and controlling behaviour;
  • non-molestation orders;
  • occupation orders;
  • financial remedy disclosure;
  • child arrangements disputes;
  • parental responsibility;
  • relocation;
  • mental health issues;
  • substance misuse allegations;
  • vulnerable parties;
  • trauma and fear;
  • children’s wishes and feelings.

This makes the use of AI much more delicate.

AI may help organise evidence, but it must not distort evidence.

AI may help create a chronology, but it must not miss safeguarding patterns.

AI may help a litigant in person draft a position statement, but it must not invent legal arguments or overstate allegations.

AI may help explain the court process, but it must not give false confidence to someone in a high-risk situation.

Family law warning:

A small debt claim and a private children case are not the same kind of legal problem. AI that is appropriate for structured debt recovery may not be appropriate for cases involving children, abuse, safeguarding and welfare decisions unless the safeguards are much stronger.

AI can help with documents. It cannot replace judgment.

The most dangerous version of the AI debate is the simplistic one.

Either AI is going to replace lawyers entirely, or AI must be resisted because it is unsafe.

Neither position is sensible.

The real issue is task allocation.

Some legal work is repetitive, structured and document-heavy. Some legal work is strategic, emotional, forensic and judgment-based.

AI may be very useful for:

  • creating first-draft chronologies;
  • summarising long documents;
  • checking whether a document answers required questions;
  • organising correspondence;
  • identifying missing dates;
  • producing task lists;
  • turning messy notes into structured drafts;
  • explaining basic procedural steps;
  • supporting fixed-fee or limited-scope legal help.

AI should not be trusted to:

  • invent facts;
  • generate witness evidence;
  • assess domestic abuse risk without human oversight;
  • decide what is in a child’s welfare;
  • replace legal advice in complex or high-risk cases;
  • cite authorities that have not been checked;
  • make safeguarding decisions;
  • tell a vulnerable person that a case is safe or hopeless without professional review.
The future is not “AI instead of lawyers”. The better future is lawyers, courts and legal support services using AI to reduce cost and chaos while keeping professional judgment firmly human.

What this means for litigants in person

Litigants in person are already using AI.

They are using it to draft emails, summarise orders, write statements, prepare questions, understand procedure, create chronologies and respond to solicitors.

Some of that is useful. Some of it is risky.

The Garfield AI development shows something important: if AI is going to be used by the public, it is better for people to have access to structured, regulated, properly designed legal tools than to be left alone with public chatbots and no legal guidance.

That is especially true in family law.

A litigant in person in family court may be frightened, traumatised, overwhelmed, neurodivergent, financially vulnerable or facing an ex-partner who is using the court process as a form of post-separation control.

They may need help to make sense of the process, but they may not be able to afford full representation.

The opportunity for family justice:

AI-assisted legal support could help litigants in person prepare more clearly, understand orders, organise evidence and reduce procedural mistakes. But it must be designed carefully around safeguarding, trauma, confidentiality and realistic legal limits.

What lawyers should take from this

Lawyers should not laugh this off.

They should also not panic.

The correct response is to understand what is happening and adapt intelligently.

The Garfield AI case is a sign that clients will increasingly expect legal services to be:

  • more affordable;
  • more transparent;
  • faster;
  • more digitally accessible;
  • more outcome-focused;
  • less dependent on open-ended hourly billing;
  • better at using technology to reduce unnecessary process.

That does not mean lawyers become irrelevant.

It means lawyers need to be clearer about where they add value.

In family law, value is not simply drafting a document.

Value is:

  • knowing what matters;
  • spotting what is missing;
  • understanding risk;
  • identifying safeguarding issues;
  • protecting the client from procedural mistakes;
  • challenging poor evidence;
  • understanding the emotional dynamics of the case;
  • preparing a realistic strategy;
  • keeping the child’s welfare central;
  • helping a client make decisions under pressure.

What regulators and courts need to think about

If regulated AI legal services expand, regulators and courts will need to keep asking difficult questions.

Those questions include:

  1. How is accuracy checked?
  2. Who is responsible for the output?
  3. How are hallucinations prevented or detected?
  4. How is confidential client data protected?
  5. When should AI use be disclosed?
  6. What happens if AI-generated material misleads the court?
  7. How are vulnerable clients protected?
  8. How are conflicts of interest managed?
  9. How are clients told what the service can and cannot do?
  10. How do courts deal with AI-assisted documents filed by litigants in person?

These questions are not a reason to stop innovation.

They are a reason to govern it properly.

A practical traffic light for AI in family law

Use of AI Risk level Family law example Practical approach
Formatting and readability Lower risk Improving layout, grammar or headings in a position statement. Useful, but still review before filing or sending.
Chronology preparation Medium risk Turning messy notes into date order. Check dates, context and missing events carefully.
Summarising evidence Medium/high risk Summarising WhatsApp messages, police disclosure or school records. Use only with careful source checking. AI may miss nuance.
Witness evidence High risk Writing or rewriting a witness statement. Do not let AI invent, embellish or reshape the client’s factual evidence.
Safeguarding or welfare analysis High risk Assessing domestic abuse, coercive control or child welfare risk. Requires professional human judgment. AI should not decide risk.

The JSH Law view

This case should be welcomed, but not misunderstood.

It shows that AI can play a serious role in widening access to legal support, particularly where the cost of traditional representation makes it irrational to bring a valid claim.

It also shows why regulation matters. The difference between a structured, regulated AI legal service and a public chatbot is not cosmetic. It is fundamental.

But family law must be approached with particular care.

The family court is not simply a debt recovery process. It deals with children, safety, abuse, fear, money, housing, contact, care and family life. That means AI tools must be built and used with much stronger safeguards.

The bottom line:

Garfield AI’s reported court success is a legal technology milestone, not a magic replacement for lawyers. The lesson for family justice is clear: AI may help reduce cost and improve preparation, but professional judgment, safeguarding awareness and human accountability remain essential.

Sources and further reading

Need help preparing family court documents or organising evidence?

JSH Law helps litigants in person and family law clients turn overwhelming paperwork into clear, structured, court-ready material.

Support can include document organisation, chronologies, evidence analysis, statement preparation, issue mapping, bundle preparation and practical case planning.

Contact JSH Law or book a consultation through the website.

This article is for general information only and is not legal advice. AI, legal technology regulation and court practice are developing quickly. Always check the latest professional guidance, court rules and regulatory requirements before relying on AI-assisted legal work.

“`