Government should set up national AI office to deal with 'harms' being caused to Irish society

Several of the Dáil committee's recommendations focused on the use of AI in systems used by social media firms, in particular how AI could be used to send children down 'rabbit holes' of negative content.

Ireland should take an “immediate, coordinated and all-of-Government approach” to artificial intelligence and establish the first national AI office next year, an Oireachtas committee has recommended.

In its first interim report, the Joint Committee on Artificial Intelligence said such an office must root out the harms being caused by AI in Irish society as part of 85 recommendations it has made to Government around the nascent technology.

“The national AI office should have well-resourced advisory panels in place for routinely collaborating with young people, older people and disabled people and including representation from those groups and also civil society groups who can provide routine expertise and insights into AI harms and benefits,” it said.

“[We should also] develop a national AI risk register within the national AI office to identify and monitor systemic risks across sectors.” 

To date, the committee has held meetings featuring speakers from a range of backgrounds, such as groups representing young people, older people, disability advocates, Traveller organisations and AI experts.

Several of the recommendations focused on the use of AI in systems used by social media firms, in particular how AI could be used to send children down “rabbit holes” of negative content.

The committee said harmful and hateful content pushed by recommender systems that use AI “needs to be better addressed in the EU AI Act [and] any codes that stem from it”.

It said: “Recommender systems should be switched off by default and social media companies should be banned from turning on recommender algorithms for accounts used by children.

“Include obligations on platform owners to prevent the use of AI-driven recommender systems for misinformation campaigns aimed at destabilising society.”

The committee noted a submission from the group CyberSafeKids, which highlighted cases taken against AI companies by the families of children who had tragically taken their own lives.

It said: “Addressing this, Dr Patricia Scanlon, Ireland’s AI ambassador, and chair of the AI Advisory Council, observed we have ‘failed miserably’, that the EU AI Act is lacking on this issue and that political will is required.

“The committee agrees and has therefore recommended this is addressed in [legislation]. It is imperative that policy makers act now to ensure young people’s protection online, including from harms associated with AI.” 

The recommendations covered a wide range of topics, and said there was a need for a national conversation around AI that had been highlighted by witnesses, as it has landed in our homes and on our devices with seemingly little information or consent.

The report referred to a “wild west” in the current AI landscape, and urged the Irish public to have information and a say in how this technology is adopted in Irish society.

In the report, committee chair Malcolm Byrne added the committee would continue to publish interim recommendations ahead of a final report, and the issues it had raised should be subject to a full debate in the Dáil and Seanad.