Tuff: Declaration would be overkill
Making litigators declare that they did not use artificial intelligence (AI) in preparing witness statements would “be counter to the pro-innovation approach adopted in the UK and reduce the efficiency that AI has introduced in certain tasks”, the Association of Personal Injury Lawyers (APIL) has argued.
APIL also warned of “unintended consequences, including satellite litigation, which would inevitably increase costs and delays”.
A Civil Justice Council (CJC) working party is consulting on introducing declarations [1] on the use of AI for witness statements prepared for trial.
For other documents – such as non-trial witness statements, statements of case and skeleton arguments – all that would be required was the name of the legal representative taking professional responsibility for them.
APIL said it agreed with the premise that witness statements for trial should always be in the witness’s own words and AI use “should be limited to basic functions such as correcting grammar or spelling” and should not be used “to change the content, the intention and the meaning” of the original statement.
“Whilst we recognise there is a risk that AI can essentially redraft a claimant’s entire words, we would be concerned with the introduction of a rule requiring a declaration that AI has not been used to generate content.
“The documents that witnesses are relying upon to write the statement may have been generated through the use of AI, and there is a risk that challenges arise if the use of AI is detected in the statement.”
In its response to the consultation, APIL said the issue should be addressed in guidance instead of formal rules.
“The guidance should reinforce the principle that witness statements should always be drafted in the witnesses’ own words, clarify the position on reliance on documents that may have been AI-generated and remind witnesses that AI could change the meaning and tone of the statement.”
On the further issue of whether there should be rules on the use of AI by human translators, APIL said: “Our view is that where the translator has signed a statement of accuracy and taken responsibility for the translation, there is no need for further rules.”
APIL said that “from a cost-saving perspective”, a publicly available AI translation tool could be useful, “provided that both parties have access to it and are able to check the translation themselves”.
However, it agreed with the proposal that the provisions for statements of truth used by experts should be amended to add a further requirement confirming that the expert’s report identified and explained any AI which had been used, other than for administrative uses such as transcription.
“We believe that guidance on what is classed as substantial generative use will be fundamental to ensure that the approach is consistent between experts in their declarations.
“Transparency regarding AI use is key for legal representatives to assess the reliability of the reports before trial, given that this falls outside their expertise.”
APIL said guidance on the distinction between “administrative and generative use” of AI would be “extremely important, as this can impact the course of a claim and legal outcomes”.
Matthew Tuff, president of APIL, commented: “A specific declaration to make it clear that a legal representative has used AI would be overkill, and undo the efficiencies that AI can offer.
“The principles of professional standards and regulations already hold the lawyer accountable for the documents submitted to the court.”