In an interesting tie-in with the Regulatory Risk session that preceded it, the answer provided by a high proportion of delegates attending the conference was that their biggest AI challenge was reconfiguring their governance policies and processes to take account of the emergence of AI into the business landscape.

There are two schools of thought in AI thinking: that AI is dangerous; or that AI is inevitable – ‘use it or die’. Zimmerman presented a variety of AI produced texts, which illustrated with a mix of humour and an underlying seriousness, both the extraordinary abilities of AI, and its inherent limitations – at least in current iterations.

While it could summarise the seminal legal case of Donoghue v Stevenson in Shakespearian rhyming couplets, its more prosaic summary of the SaamCo case, while superficially eloquent and persuasive, hid a silent threat - the omission of any reference to the Manchester Building Society case. As more than one US attorney has discovered to their cost, AI has both problems of omission and commission. Unchecked, it can be a commensurate liar.

AI can appear seductively omnipotent – leading to a false sense of security and the real danger of overreliance. The risk of overuse and misuse has to be balanced with the likelihood, that, in some circumstances, failing to use AI tools may itself be regarded as negligent.

Key takeaways from the session

• Always manually check AI output – for accuracy and completeness.
• Use it as a tool to assist your thinking and output, but do not rely on it.
• Understand that AI is not intelligent and cannot reason.
• Choose your AI tools carefully, and their use cases. Using AI to generate legal research and legal advice is likely to be highly risky.
• Ensure your clients understand how you are using AI, and give informed consent – particularly in situations where they are actively seeking to reduce time and cost by usage of AI tools.

For more information on AI risks, read Miller’s whitepaper: ‘The use of AI in legal services – opportunities, risks and mitigation’ below produced in conjunction with Simmons & Simmons - and look out for our upcoming webinar Integrating AI into Legal Practice.