Government departments deny using AI to answer questions from TDs or the media

There has been a marked increase in the number of parliamentary questions asked by TDs this year — some 50,000 questions up to the end of July, compared with just over 40,400 in the same period last year — prompting focus on the use of AI by politicians to submit questions.

Several Government departments insist they do not use artificial intelligence (AI) to answer questions from TDs or the media.

There has been a marked increase in the number of parliamentary questions asked by TDs this year — some 50,000 questions up to the end of July, compared with just over 40,400 in the same period last year — prompting focus on the use of AI by politicians to submit questions.

The Houses of the Oireachtas have recently confirmed questions will be submitted using its own in-house system rather than email from September, but departments have said while civil servants can use certain AI tools, they remain forbidden from using AI systems to generate answers. 

Most departments said their advice was based on guidance received from the National Cyber Security Centre (NCSC).

The Department of Enterprise, Tourism and Employment said it was "committed to harnessing the benefits of AI to enhance service delivery, while ensuring its use is grounded in a human-first approach that underpins public trust and protects people’s rights". 

It said it had told staff in 2023 that generative AI could deliver efficiencies in some areas "but also set out the risks of using such technologies".

The department said its "general principle" was the technology could be used in most situations where it could improve the efficiency of its work and functions, but staff are advised they must "make their own value judgments and be transparent and vigilant about any significant use of generative AI" and remain "100% responsible for all outputs".

The Department of Transport said it used a “human in the loop” principle with all use of AI, which means people oversee what AI does.

"No decisions are taken by AI and, where AI is used in work, staff members remain fully responsible for any output that may have involved AI tools."

The Department of Justice said it "does not currently use generative AI tools for answering media queries and parliamentary questions" and it "continues to assess the opportunities presented by all digital tools, including AI, to improve the delivery, security, efficiency and accessibility of services to our staff and service users".

"Approval for adoption of any new tools or solutions within the department would be subject to appropriate governance processes, including risk assessment."

The Department of Further Education said staff had been given a tool "to use generative AI to assist with activities such as summarising content, explaining information, and proposing draft responses to general informational queries". However, it said, where generative AI is used, "it is used only as an assistive tool" and "does not replace existing review, clearance, approval, or accountability arrangements".

The Departments of Taoiseach, Children, Health, Housing, Agriculture, Social Protection, Culture, Rural Affairs and Foreign Affairs said they adhered to the Government guidelines for responsible AI use in the public sector published last year. 

These guidelines provide the principles that all departments and public bodies should adhere to when considering how AI can be used.