The Government must prioritise “robust safeguards” against harms posed by artificial intelligence (AI), according to a human rights body that warned of the “high risks” carried by the use of AI in public services.
Liam Herrick, chief commissioner of the Irish Human Rights and Equality Commission (IHREC), said the public had a strong appetite for effective regulation.
“We must be wary of a self-interested, anti-regulation discourse which would strip away fundamental rights protection in the interests of corporate profits,” he said.
He was speaking to Oireachtas committee members considering the use of AI in the public service and by State bodies.
READ MORE
Mr Herrick told committee members the use of AI in public services carried high risks, “especially when used to make decisions about essential entitlements and supports”.
Rebecca Keatinge, head of monitoring and compliance at the IHREC, said there have been “arresting” examples of “algorithmic discrimination” abroad when used in public services.
For example, she noted the “infamous child benefit scandal” in the Netherlands in 2018, which saw tax authorities use a “highly discriminatory” algorithm in an effort to identify fraud within the child benefit system.
The algorithm, which she said had its own decision-making capacity, used citizenship as one of the risk factors and led to “whole nationalities being essentially blacklisted for very small administrative errors in their applications”.
“It led to really devastating human impacts,” she told committee members, adding that tens of thousands of people were affected.
Denmark, which has a high level of social welfare provision, also employs different algorithms to help manage its social welfare system, including fraud control, she said.
Alongside using citizenship as an indicator, it also picks out “atypical characteristics which don’t essentially concur with Danish social norms,” she said, such as household size.
Ms Keatinge said this could constitute “social scoring”.
Dr Kris Shrishak, senior fellow at the Irish Council for Civil Liberties (ICCL), meanwhile, urged the committee to push for an independent National AI Office, with a dedicated budget, commissioner and technical experts.
A National AI Office is set to be established by August 2026 and will act as a focal point for AI in Ireland, “encompassing regulation, innovation and deployment,” the Department of Enterprise has said.
[ Government working on plan to use AI and apps to solve public sector issuesOpens in new window ]
“It should not be housed within any of the Government departments,” Dr Shrishak told committee members.
Dr Shrishak, who leads the ICCL’s work on the AI Act, noted the use of three separate AI chatbots by the Department of Justice to date for queries relating to immigration and international protection.
He told committee members that through freedom of information requests, it was discovered the department “did not run a tender process, performed no risk assessment, no bias tests, and no environmental impact assessment for any of these”.
