Oireachtas committee members 'sought to water down report on artificial intelligence'

Oireachtas committee is finalising its third interim report on AI

Oireachtas artificial intelligence committee chairman Malcolm Byrne proposed a number of amendments to a draft report. File picture: Brian Lawless/PA

A disagreement has emerged at the Oireachtas artificial intelligence (AI) committee, with some members claiming that others are attempting to “water down” the committee’s opinions regarding AI corporations and regulation.

The committee is currently finalising its third interim report regarding AI in the Irish context, covering issues including the fallout from revelations that Grok, the artificial intelligence (AI) chatbot of social network X, had been used to digitally “undress” people using existing photographs.

Changes suggested by Malcolm Byrne

At a meeting to discuss proposed amendments to the report, committee chair Malcolm Byrne suggested the draft report be updated to remove an assertion that the self-administered risk assessments favoured by AI companies “would not be accepted in any other industry”, to be replaced with a paragraph stating “that AI should be subject to more demanding assessment before deployment”.

Mr Byrne further suggested that a statement in the draft report — asking why if medical devices must be “thoroughly vetted” before reaching consumers then “why are we not doing that with AI” — be replaced with a paragraph asserting that “the relevant question is which assessment requirements are justified by the risks of particular AI applications”.

Grok provided sexualised images of children

Separately, the draft report deals with a meeting last January with AI junior minister Niamh Smyth regarding evidence that Grok had been used by X members to digitally manipulate photographs of real people, including children, to produce sexualised images.

At that meeting, People Before Profit TD Paul Murphy indicated to Ms Smyth that she had been misled by X regarding assertions that Grok had been blocked from creating sexual imagery from photos, a point he proved by declaring that one of his staff had been able to digitally undress him, with his permission.

In the initial draft report, referring to the same meeting, Fine Gael TD James Geoghegan suggested any reference to the interaction between the minister of state and Mr Murphy should be removed, with any mention of the incident to be limited by reference to the transcript.

At the meeting of committee members this week, it was agreed to leave in the reference to the incident however, and to instead include a link to the transcript within the paragraph under discussion.

What's your view on this issue?

You can tell us here

Of the deliberation underway to compile the report, Mr Murphy said that some Government TDs “have been seeking to water down the draft report”.

“At a time when we are seeing a major push to undermine regulation of Big Tech and AI at a European level, it is vital that we accurately reflect the stark evidence put before our committee,” Mr Murphy said.

AI represents a dangerous threat to democracy as well as our environment.

"As a committee, we should be standing up in defence of fundamental rights, not the profits of Big Tech corporations.”

However, committee chair Mr Byrne said that there had been no “disagreement” between members. 

“I actually think we strengthened some of the recommendations around criticisms of Grok. I don’t think there was any disagreement there,” Mr Byrne said.

He added that, in his opinion, the only “rows” had emerged from discussions regarding a potential windfall tax on tech companies and on the digital omnibus — a simplification measure for the EU’s digital laws proposed by the European Commission last November, aiming to streamline the laws.