Judiciary to issue detailed practice note on AI
The judiciary is developing a detailed practice note to guide practitioners and parties on the use of AI in litigation, Ms Justice Eileen Roberts has said.
The High Court judge was chairing the Dublin Bar Solicitors Association (DSBA) webinar, Legal Perspectives on Generative AI – From Legal Privilege to Contracts to Publication & Defamation (28 April).
Referencing Guerin v O’Doherty, the judge said “For some time, we watched in the Irish courts as these hallucinations featured around the world, and unfortunately, we've now been victim of them ourselves.”
In light of this, the judiciary, working with solicitors and barristers, will produce “a detailed practice note which will help practitioners and parties appearing in the court to understand what is expected of them”.
Technology evolves quickly, more so than is possible for a legal system that “never legislated for or intended or knew about Generative AI”, Ms Justice Eileen Roberts commented.
Although existing frameworks such as the EU AI Act and General Data Protection Regulation provide context, the system will remain “behind the curve” and must be “ready to change”, she added.
The courts accept that AI use is widespread and are “clear in Guerin that there is nothing objectionable” in that.
She added that requiring disclosure of AI use “every single time” could create “a separate piece of satellite litigation and inquiry as to what exactly has been done”.
“We don't want to create a situation where the disclosure of AI itself becomes a new ground of debate and argument between the parties.”
Accurate and correct
The central judicial concern is to ensure that material before the court is “accurate, that they are correct, that we are getting information that is not misleading us”.
Ensuring that witness statements and expert reports are not adversely influenced by the use of AI is a key concern.
Overuse in victim statements
Ms Justice Roberts added: “Witness statements are a thing where I'm where I am feeling that AI is being used, probably a little more than might be helpful.”
She added: “There is definitely a view that unrepresented parties in particular are relying increasingly on AI.
“We haven't yet had a case where there has been a claim for privilege, for example, in relation to documents generated in that way, but it may well arrive.”
Accountability and scrutiny are vital, she said.
Truthfully and honestly
“If you're using AI,” Judge Roberts said, “you should be able to answer the question if asked as to what you've used it for”, and to do so “truthfully and honestly”.
And, if an authority is relied upon, it should be produced as “a very easy way of at least confirming the authority exists”.
New categories of dispute are anticipated, including defamation and intellectual property claims, and both judiciary and practitioners are “trying to navigate what is entirely new for all of us”.
Ms Justice Roberts noted that risks extend beyond practitioners where clients using public AI tools to “double check everything”, might potentially create further legal and confidentiality issues.
“So hopefully over the next short while, we will, after engagement with both branches of the professions, come up with a useful practice note,” Ms Justice Roberts said.