Regulator Watch
What BCI and SRA have said about AI — and what they have not said yet. A summary of regulatory positions, existing obligations, a side-by-side comparison, and a practical compliance checklist for firms and chambers.
What BCI and SRA Have Said About AI — And What They Haven’t Said Yet
Overview
Neither the Bar Council of India (BCI) nor the Solicitors Regulation Authority (SRA) has banned the use of AI by legal practitioners. Both have made clear — in different ways and to different degrees of specificity — that existing professional obligations apply in full to any work involving AI tools.
This guide summarises what each regulator has said, identifies the gaps, and sets out a practical compliance approach for firms and chambers.
SRA Position (England & Wales)
Key Statements
| Date | Document | Key Point |
|---|---|---|
| Nov 2023 | SRA Statement on AI | AI use is permitted. Professional obligations — competence, confidentiality, independence — apply in full. |
| 2024 | SRA Thematic Review (Risk Outlook) | Supervision of AI output flagged as a key risk, particularly for smaller firms. Training and oversight gaps identified. |
| Ongoing | SRA Standards and Regulations | Rule 1.4 (competence), Rule 6.3 (confidentiality), and the Solicitors Act 1974 all apply regardless of the tools used. |
What the SRA Has NOT Said
- It has not issued prescriptive rules on which AI tools may be used.
- It has not specified what a compliant AI use policy must contain.
- It has not created an ‘AI defence’ for disciplinary matters.
- It has not defined what level of AI-output review constitutes reasonable supervision.
The SRA’s approach is principles-based: if something goes wrong in AI-assisted work, the question will be whether the solicitor discharged their professional obligations — not whether the AI was at fault.
Bar Council of India Position
Current Status
As at the date of this module, the BCI has not issued specific guidance on the use of AI by advocates. There is no BCI circular, notification, or resolution specifically addressing AI-assisted legal work.
However, the following existing rules and statutes apply and govern AI use by default:
| Rule / Statute | Relevance to AI Use |
|---|---|
| Advocates Act 1961, S.49 (Standards of professional conduct) | Any tool that produces work an advocate presents as their own must meet professional standards. |
| BCI Rules, Part VI, Chapter II (Standards of Professional Conduct) | Duty to client includes competent advice — AI-assisted advice is not exempt. |
| BCI Rules on confidentiality | Obligation not to disclose client information extends to third-party AI tools. |
| Indian Contract Act / fiduciary duties | Advocates owe fiduciary duties that are not delegable to automated systems. |
What Is Expected
Given the BCI’s historical approach to technology and professional conduct, future guidance is likely to emphasise:
- Personal accountability — the advocate remains responsible for all work presented to courts or clients.
- Non-disclosure — client information may not be shared with third parties (including AI platforms) without consent.
- Competence — advocates must understand the tools they use well enough to identify errors.
- Court duties — misrepresenting AI-generated content as original research is a contempt risk.
Side-by-Side Comparison
| Issue | SRA (England & Wales) | BCI (India) |
|---|---|---|
| Specific AI guidance issued? | Yes (2023, thematic review 2024) | No — general conduct rules apply |
| AI use permitted? | Yes, with professional obligations | Yes, subject to existing rules |
| Competence obligation applies? | Yes — Rule 1.4 | Yes — Advocates Act / BCI Rules |
| Confidentiality obligation applies? | Yes — Rule 6.3 | Yes — BCI Rules Part VI |
| Supervision requirement? | Yes — flagged in thematic review | Implied by personal responsibility rules |
| Documentation required? | Recommended (thematic review) | No specific requirement yet |
| Sanctions for AI misuse? | Yes — conduct matters under SRA Code | Yes — professional misconduct under Advocates Act |
Practical Compliance Checklist
Firms and chambers operating under either or both regulatory regimes should consider:
Supervision
- AI-generated work product is reviewed by a qualified professional before it leaves the firm
- The reviewer has sufficient knowledge of the relevant law to identify errors
- Junior staff using AI have clear guidance on when to escalate for review
Documentation
- AI use is recorded in the client file where it materially contributes to a work product
- The firm maintains a log of which AI tools are approved and how they are used
- Client-facing documents do not misrepresent AI-drafted content as independently authored
Training
- All fee-earners have received basic training on AI limitations (hallucination, knowledge cutoff)
- All fee-earners understand the firm’s data classification and which data may be input to AI
- Training is updated annually as the regulatory and tool landscape evolves
Policy
- The firm has a written AI use policy that has been communicated to all staff
- The policy identifies approved tools and prohibited uses
- The policy is reviewed at least annually
What to Watch
The regulatory landscape is moving. Practitioners should monitor:
- SRA — Further thematic review outputs; possible prescriptive guidance on supervision standards and AI logging.
- BCI — The first AI-specific circular or notification, likely following a high-profile incident of AI misuse in Indian courts.
- Courts — Practice directions in High Courts and tribunals — several have already issued or are developing rules on AI-generated submissions.
- DPDP Rules — Subordinate legislation under the DPDP Act 2023 may impose additional obligations on firms processing personal data via AI platforms.
Part of the AI Foundations for Lawyers series.