Tanna: High-level guidance
The Society for Computers and Law (SCL) has launched a set of free contractual clauses for transactions involving artificial intelligence (AI).
Minesh Tanna, chair of the SCL’s AI group, said the impact of AI on contracts was “not straightforward, particularly at this early relatively early stage in AI adoption”.
The clauses, drafted by members of the SCL’s AI group over several months, are intended to provide guidance on the types of issues that should be addressed in contracts for transactions involving AI, and are not to be relied on as legal advice.
The 60 pages of clauses cover areas such as primary obligations, intellectual property/licensing and contractual protection, from the point of view of both customer and supplier.
They also include drafting notes to help explain the rationale for the positions taken, the drafting itself and other points to consider, along with suggested contractual definitions for key AI concepts.
Mr Tanna, partner and global AI lead at City firm Simmons & Simmons, said the clauses covered procurement of a ready-made AI solution or a bespoke arrangement where the customer helped the supplier develop it.
“Due to the way in which AI models are created, where the customer has a say, the contractual approach varies.
“If a contract is signed at the start of the development process, the customer should be able to be more prescriptive about training and testing. If the product is already developed, the customer will have to rely on promises about how it has been trained and tested.”
Mr Tanna said the clauses should be treated as “high-level guidance” for AI contracts, intended to draw attention to the ways in which AI is likely to have an impact and raise awareness.
He said generative AI heightened some of the risks associated with the technology, such as inaccuracies, data privacy and intellectual property.
“Where a contract relates to generative AI, the parties should be made particularly aware of how the risks play out.”
Mr Tanna said the AI group had received “lots of positive feedback” on the clauses and planned to create a process for gathering formal feedback early next year, which would then be fed back into an updated version.
The other element that he wanted to incorporate was the forthcoming EU AI Act, the bloc’s first regulatory framework for AI, which could be agreed as early as this month but was more likely by the end of the year, with the final text emerging early next year.
The Act will have extra-territorial effect where businesses outside the EU sell systems to businesses inside the EU or make them available to consumers there.
Members of the AI group’s committee include solicitors at City firms, in-house counsel and legal academics.
The group aims, among other things, to provide a discussion forum for those interested in legal, regulatory and ethical issues relating to AI, increase knowledge and awareness of AI among lawyers and contribute to policy and legal developments.
Leave a Comment