AI emotion-recognition technology will be unacceptable in any new AI framework, Dara Calleary (Minister of State for Trade Promotion, Digital and Company Regulation) has said.

Speaking to Eolas magazine, the minister said that he does not want people replaced by AI but he wants their skills to be enhanced.

The minister said that AI will change the way we work but that the EU AI Act has designated areas that are high-, medium- and low risk.

Health and law enforcement will be classed as high-risk with stringent compliance requirements, he said.

Transparency requirements

Medium risks are transparency requirements while low risk means minimal if any restrictions.

Healthcare, education and general government services will all be impacted by the rise of AI, he said, as part of a general culture of innovation, but citizen trust is essential.

Subject to EU Parliament agreement, the AI Act will enter into force in early summer, he said, with three further years get it into place.

“We will be doing an awful lot of it in the first 18 months. It builds the guardrails, it is accessible, and people will know what is allowed in terms of what is high risk, what is medium risk, what is low risk, and what is unacceptable in terms of the technology,” he said.

Undesirable outcomes

The minister said he was conscious of the potential for AI to produce or amplify inequitable or undesirable outcomes, for individuals and society.

“I want AI to be used to break down barriers,” he said and that is why he is working closely on the public service impact.

“I see it making the delivery of healthcare much fairer. When you see the waiting lists for healthcare, it is people on lower incomes that are on the waiting list or waiting longer.

“I see using AI and getting access to technologies as having the potential make it far more equal. AI technologies have the potential, if used properly, and there is a strategy in place like ours, to open up all sorts of opportunities for every citizen,” he said.

The minister said that a planned public information awareness campaign will explain what AI can do.

“People are using it every day. I think that is one thing that people forget about. If you engage with a chatbot, that is AI,” he said.

“Every party should lay out their stall on AI and how they see it being utilised over the course of the next five years.”

Rapid evolution

The minister said the AI Advisory Council will grapple with the rapid evolution of AI, and that an AI committee is likely in the Oireachtas.

In January 2024, the AI Advisory Council met for the first time, after 400 applications came in to be part of it, the minister said.

“There is an extraordinary level of talent on AI within the country,” he told Eolas.

“I would ask policymakers to listen to what the Advisory Council comes up with and to engage with the work it is going to produce. I see it as having an ESRI-style role, without the budget, in terms of advice and issues,” he said.

“I want them to demystify it and come back to the principles: is it trustworthy; is it person-centred; is it ethical?” he said.