JSH LAW · EVIDENCE · ACCOUNTABILITY · ENGLAND & WALES
Evidence, Accountability and the Right to a Fair Hearing
Artificial intelligence can help organise a family-court case. It can also introduce a false authority, distort a witness’s account or make a disputed allegation look like an established fact. The question is what people can safely rely on—and how they can understand, check and challenge material that may affect a child or family.
This resource centre brings together JSH Law’s published guides and commentary on AI, digital evidence, professional assessments, safeguarding and legal responsibility. Start with the problem in front of you, then follow the sources and the procedure that apply.
For parents, litigants in person, practitioners and researchers. Published resources selected and links checked: 15 October 2026. Individual articles have their own dates and review limits.
The quick answer
Use AI to support preparation with care. Keep the original evidence, verify legal references and preserve the difference between fact, allegation, opinion and judicial finding. A polished output is a starting point for checking, not a substitute for it.
If a hearing or deadline is close, start with the latest sealed order, the issue the court must decide and the particular material you need to address. These resources do not change a deadline or replace individual advice.
Start with the problem you need to solve
- A case citation or legal answer looks wrong: start with legal research and responsibility.
- Messages, images or recordings are disputed: use the digital-evidence guides.
- Monitoring, impersonation or message patterns concern you: read technology-facilitated abuse and safeguarding.
- A report relies on information you cannot trace: start with assessments, records and fair challenge.
- You want to use AI for preparation: begin with responsible tools and case organisation, then check privacy and permissions.
AI in Family Courts and judicial decision-making
A useful AI summary is not the same thing as a judicial reason. These resources examine who evaluates the evidence, how meaningful oversight can be demonstrated and whether the person affected can understand the reasoning. The distinction matters especially when decisions concern a child’s safety, care or relationships.
The judiciary’s position statement of 9 October 2026 says judicial decisions remain with judicial office holders: AI must not displace reasoning, evaluation of evidence or discretion. This is judicial governance material; it does not establish that any particular tool was used in your case.
PUBLISHED · CORE READING
AI Is Already in the Courts: Why Family Justice Needs Hard Red Lines
A family-justice argument for accountable judicial AI: evidence integrity, independent reasoning, children’s welfare and meaningful human review. Read the proposed safeguards as commentary, alongside the later official position below.
PUBLISHED · CORE READING
A Human in the Loop Is Not Enough: “Humanwashing”, AI and the Future of Judicial Decision-Making
Why a human signature does not, by itself, demonstrate meaningful oversight. This analysis examines research on perceived fairness; it does not establish that a particular Family Court uses algorithmic adjudication.
PUBLISHED · CORE READING
AI in Justice Must Be People-Centred — Not Just Another Tool to Process Vulnerable People Faster
An accessible starting point on participation, digital exclusion, safeguarding and the difference between processing cases faster and helping people obtain justice.
AI, legal research and professional responsibility
A plausible citation may be invented. A real judgment may be used for a proposition it does not support. Start with verification and the integrity of the final document; then examine the distinct duties and possible consequences for the person who prepared or adopted it.
PUBLISHED · CORE READING
The Legal AI Trust Test: Seven Checks Before You Rely on ChatGPT or Any AI in Court
Seven checks for an AI-assisted legal answer or court document: a practical route back to the original law, source evidence, procedure and human review.
PUBLISHED · CORE READING
AI Slop, Costs and Litigants in Person: Should Courts Sanction Bad AI Use?
A considered discussion of poor AI-assisted filings and costs. It distinguishes the access-to-justice problem from irresponsible use; procedural consequences depend on the applicable jurisdiction and facts.
PUBLISHED · CORE READING
AI Can Write 300 Pages. That Does Not Make It Advocacy.
Case commentary on the difference between producing text and presenting a case. Useful for understanding relevance, verification, proportionate submissions and preserving the witness’s own evidence.
Digital evidence, deepfakes and authentication
Messages, images and recordings are often central to a family case. The practical questions are where the material came from, whether it has changed, what context is missing and what it actually supports. An allegation that something is “AI” needs a focused evidential basis rather than suspicion alone.
FPR 22.1 gives the court control over the issues requiring evidence, its nature and how it is placed before the court. Preserving a file and obtaining permission to rely on it are different tasks.
PUBLISHED · CORE READING
AI Evidence in the Family Court: Deepfakes, Screenshots & Digital Evidence
The central digital-evidence guide: provenance, originals, screenshots, recordings, disputed authenticity, forensic questions and the limits of automated detection.
PUBLISHED · CORE READING
Can You Use WhatsApp Messages as Evidence in Family Court? The 2026 Rules Explained
Practical guidance on preserving conversations, retaining context, selecting relevant messages and preparing material for the court. Check current bundle requirements and your own directions.
PUBLISHED · CORE READING
Can a secret recording be used as evidence in the Family Court?
A detailed guide to covert recordings, evidential use and safeguarding risks. Reading it is not permission to record a child, a professional meeting or a court hearing.
Technology-facilitated abuse and safeguarding
Technology can preserve evidence and expose an alleged pattern. It can also enable monitoring, impersonation, intimidation and selective presentation. These guides keep personal safety, the child’s experience and the distinction between an allegation and a finding in view.
PUBLISHED · CORE READING
Technology-Facilitated Abuse in the Family Court: GPS Tracking, Smart Devices & Digital Evidence
How tracking, smart devices and digital communications can feature in abuse allegations, and why evidence preservation must be considered alongside personal safety.
PUBLISHED · CORE READING
Can AI Prove Coercive Control? No — But It Can Help the Court See the Pattern
An applied discussion of extracting possible patterns from messages while preserving source material, context and competing explanations. It does not independently validate a commercial tool or turn a classification into proof.
PUBLISHED · CORE READING
AI Can Hack Humans Too: Why Family Justice Must Defend the Human Decision-Maker
Commentary on social engineering and the manipulation of trust. It asks how family justice can protect the people who supply, assess and decide upon evidence.
Cafcass, expert assessments and procedural fairness
A report may combine direct observation, information supplied by others and professional interpretation. The route to a useful response begins by tracing those layers and explaining why a disputed point matters. These resources provide that foundation without assuming that a particular organisation uses AI.
For the wider records question, use Family Court Disclosure: How to Obtain Police, Medical, School, Social Care & Cafcass Records. Part 21 addresses disclosure and inspection, while Part 25 governs experts and assessors. The correct route depends on the proceedings and what is sought.
PUBLISHED · CORE READING
How to Challenge an Inaccurate Social Worker Report: A Line-by-Line Guide
A substantial resource for separating factual errors, missing context, professional opinion and unsupported reasoning. Includes worked examples and a structured response matrix.
PUBLISHED · CORE READING
How to Obtain Cafcass Records and Case Notes: Subject Access Requests and Family Court Disclosure
A practical guide to the distinct routes for personal-data access and court disclosure, including Cafcass England, Cafcass Cymru and children’s information.
PUBLISHED · CORE READING
Behaviour Is Not Need: Why Family Lawyers Need Cross-Disciplinary Literacy
Assessment literacy for family proceedings: how to question the connection between observed behaviour, underlying need, professional interpretation and the proposed plan.
AI regulation, transparency and legal accountability
Responsibility depends on the actor, task, data and applicable legal duty. Useful governance asks what a system may do, what must be checked and who remains accountable. A policy proposal, a regulator’s guidance and a procedural rule do not have the same legal status.
PUBLISHED · CORE READING
AI Liability in UK Legal Practice: SRA Duties, Court Evidence and Data Protection
The existing technical starting point on professional accountability, confidentiality, court evidence, data protection and liability. A future expansion will deepen the family-practice analysis.
PUBLISHED · CORE READING
Before You Press Submit: Why AI Prompt Governance Matters in Legal Work and Family Justice
Practical governance before information is entered into an AI system: task suitability, sensitive inputs, output verification and the risk of changing the evidence.
PUBLISHED · CORE READING
When AI Stops Answering and Starts Acting: Agentic AI, Legal Responsibility and the Next Professional Risk
The additional questions raised when AI can access tools and take actions: authority, permissions, supervision, audit records and human responsibility.
Access to justice and responsible legal technology
Technology is useful when it helps a person understand the process, organise reliable material and participate effectively. These resources focus on a manageable case file, proportionate preparation, privacy and the continuing need for human judgment.
PUBLISHED · CORE READING
How to Use ChatGPT as a Litigant in Person: A Practical Guide to Preparing Your Court Case
A practical preparation guide for litigants in person: source files, chronologies, issues, drafts and checking. Product settings change, and an account feature does not authorise an upload.
PUBLISHED · CORE READING
Your Family Court Case Needs a System: CRM, AI and Evidence for Litigants in Person
How to organise orders, deadlines, source evidence and working analysis without allowing an AI-generated summary to become the case’s source of truth.
PUBLISHED · CORE READING
Trauma-Informed Lawyering Is Not Enough: Legal Technology Needs Guardrails Too
Why communication, vulnerability and participation matter in legal-technology design, and why helpful automation still needs evidence and safeguarding controls.
The JSH Law Six-Question Check
Use this preparation framework on one important statement, summary or recommendation at a time. Its purpose is to identify what needs checking or clarification. It is not a legal test, a risk score or a method for deciding whether somebody is telling the truth.
- Source
- Where did the information actually come from? Identify the earliest available source rather than counting repeated summaries as independent support.
- Status
- What kind of information is it: allegation, admission, observation, record, professional opinion, expert opinion or judicial finding?
- Context
- What surrounding events, qualifications, missing material or competing explanations are needed to understand it fairly?
- Participation
- Has the affected person had an appropriate opportunity to understand and respond, with relevant communication and participation needs recognised?
- Consequence
- What conclusion, recommendation, direction or order is said to follow—and what reasoning connects it to the material?
- Responsibility
- Who must do what next, by when, and how will completion, implementation or review be checked?
Read the full JSH Law Six-Question Check and worked example. Keep it distinct from the separate Legal AI Trust Test and the topic-specific checking tools linked above.
Wider published commentary
Some questions need a wider lens: the structure of legal services, the proposed legal status of autonomous systems and the principles that should guide legal technology. These pieces provide context, while the practical routes above remain centred on family justice.
- Should AI Become a Legal Person? Imogen Rivers, Robot Rights and the Responsibility Gap
Academic commentary on legal personhood and responsibility gaps. It explores proposals rather than recognising AI as a legal person under current English law. - Family Law Technologists: Where AI Meets Evidence, Safeguarding and Justice. My Two Penneth.
A broader statement of JSH Law’s approach to evidence, safeguarding, verification and responsible family-law technology. - Legal AI Is Rebuilding the Legal Services Market — But Who Will It Serve?
Wider legal-services commentary on who benefits from AI infrastructure, automation and changing business models. Useful context rather than a family-procedure guide.
Eight developing research subjects
The published resources above are the foundation for a wider programme: Artificial Intelligence, Family Justice, Evidence and Legal Accountability. All eight original subjects remain in scope. Some will deepen an existing article; others need a new technical or comparative publication.
Research programme—not a list of completed articles. The entries below describe planned work, subject to primary-source verification and editorial approval. No unpublished article is linked or presented as available.
| Subject | Approach | Contribution to develop |
|---|---|---|
| AI hallucinations in the courts | New family-focused flagship analysis | Build on the checking and costs guides, with verified distinctions between false authorities, mistake, professional failure and the separate thresholds for procedural consequences. |
| Judges and artificial intelligence | Substantial expansion of existing coverage | Develop the judicial-AI article with current governance, reasons, meaningful review and the practical safeguards relevant to family proceedings. |
| Algorithmic opacity and access to justice | New flagship analysis, with staged comparisons | Examine explainability and effective challenge. Any organisation-specific account must begin with evidence of actual deployment; a hypothetical system must remain labelled hypothetical. |
| AI-generated evidence and deepfakes | Substantial expansion of the existing evidence guide | Deepen authentication, evidential weight, proportionate directions and expert questions while retaining the established digital-evidence URL. |
| The EU AI Act and UK law | Separate specialist and comparative strand | Examine territorial scope, regulatory divergence and meaningful relevance to UK family practice. EU provisions must not be presented as the Family Procedure Rules. |
| AI in medical and expert decision-making | Staged family-evidence and specialist analysis | Begin with the basis and limits of AI-assisted expert assessments; retain wider clinical and medical-regulatory research for a distinct, properly scoped publication. |
| AI transparency and disclosure obligations | New family-litigation flagship analysis | Distinguish disclosure of source material, methods and AI involvement from access to personal data, privilege and public reporting. |
| Responsibility when AI gets it wrong | Substantial expansion of existing liability coverage | Develop actor-specific duties, supervision, causation, loss and remedies, with a family-practice focus and broader accountability research retained. |
The intended contribution is detailed, source-linked analysis that practitioners can scrutinise and individuals can understand. A website article is not peer-reviewed scholarship merely because it discusses academic work. Inaccessible journal papers and unverified leads will be identified as such rather than presented as examined sources.
Official sources and how to read them
The hub is a reading pathway, not a replacement for primary law. Distinguish statutory and procedural requirements, binding judgments within their scope, professional or judicial guidance, commentary and proposals. Foreign research can illuminate an English-law question without determining its answer.
- Family Procedure Rules Part 22: evidence—including the court’s power to control evidence under rule 22.1.
- Family Procedure Rules Part 21: disclosure and inspection—read with the procedure applicable to the case.
- Family Procedure Rules Part 25: experts and assessors—the expert framework; read it with the published FPR amendments and commencement information, rather than treating a single consolidated page as the whole current position.
- Judicial AI position statement, 9 October 2026—governance and safeguards; distinguish it from legislation and case-specific findings.
- SRA warning notice: misuse of AI—regulatory guidance for the firms and individuals within its scope.
Links and resource selection were checked on 15 October 2026. This does not re-date every linked article or certify all of its legal or technical claims. Read each article’s date, source notes and limitations, and check current primary materials before taking a procedural step.
Need help turning difficult material into a clear next step?
JSH Law offers defined, non-reserved support with evidence organisation, source-linked chronologies, court-document preparation, professional-report responses and hearing preparation. Start with the immediate problem, the latest order and the next deadline.