Emerging risk
Professional Indemnity and artificial intelligence: emerging risk for South African professionals
How the growing use of AI tools in professional services is starting to affect PI risk, disclosure and underwriting.
Key takeaways
- ✓Using AI tools in your professional work doesn't remove your professional duty of care — you remain responsible for the advice, design or output you deliver to a client
- ✓Insurers are increasingly asking about AI tool usage as part of underwriting, particularly for technology, legal, financial and consulting practices
- ✓An error originating from an AI tool is generally treated the same as any other professional error for PI purposes — your policy responds to your liability for the deliverable, not to how it was produced
- ✓Disclosing significant reliance on AI tools in client-facing work is emerging as good practice, both contractually and for accurate underwriting
Does using AI change what your PI policy covers?
Generally, no. Cover responds to your professional liability for the advice, design or service you deliver, regardless of what tools you used to produce it. Using an AI tool doesn't itself void or reduce cover, provided you exercised appropriate professional oversight and judgement over the final output before it reached your client.
Where new risk is emerging
The real exposure is over-reliance on unverified AI output without adequate professional review — a lawyer citing a fabricated case, an accountant accepting an AI-generated figure without checking it, a developer shipping AI-generated code without testing.
Insurers are starting to treat inadequate oversight of AI-assisted output as a potential aggravating factor in a negligence claim, in much the same way inadequate review of a junior staff member's work has always been treated.
What insurers are starting to ask
Expect increasing underwriting questions about whether AI tools are used in producing client deliverables, what review process exists before AI-assisted output reaches a client, and whether client data is shared with third-party AI tools — which raises a separate POPIA and Cyber consideration alongside the PI question. See our PI vs Cyber Insurance guide for that overlap.
Practical steps for professionals using AI tools
A few habits materially reduce this emerging risk:
- Maintain the same review and sign-off standard for AI-assisted work as for any other output
- Disclose material AI use to clients where it could reasonably affect their expectations
- Check whether client data entered into AI tools is handled consistently with your POPIA obligations
- Ask your insurer directly whether AI use is now a specific underwriting question at your renewal
FAQ
Frequently asked questions
Not automatically. Insurers are still developing their approach to this — the more relevant factor is whether your review and quality-control processes are adequate, AI-assisted or not.