MEPs have been urged to consider the ‘significant risks’ that next-generation Artificial Intelligence applications could entail, particularly with regards to discrimination, employment, and social exclusion.
On Thursday (14 January), members of the European Parliament’s Special Committee on Artificial Intelligence played host to experts on the subject from the European University Institute (EUI). Members were briefed on the benefits of future technologies alongside the potential pitfalls.
“AI certainly provides great opportunities for individuals and the world’s societies with regard to sustainability, and the possibility of increasing knowledge,” Dr Francesca Lagioia, Senior Research Fellow said. “But it also involves some significant risks, such as for example, unemployment, discrimination, and exclusion.”
“And so, the challenge is, how to support useful AI applications while preventing harms to individuals and society.”
Concerns surrounding the discriminatory elements of certain AI applications was a point picked up by Renew MEP Karen Melchior, who called for more transparency into the ‘black box’ operation of certain technologies.
“If we’re looking at, for example, the hiring of people and choosing which candidates to bring in and discrimination, then it becomes much more difficult to quantify what is going on,” Melchior said, adding that Parliamentarians should look for clarity on where they should be “applying the requirements on the development of algorithms so that we open up what is often seen as a black box.”
Parliament’s probe into the potential pitfalls of AI comes at a time when the European Commission is readying legislation to follow up on its 2020 White Paper on AI, which laid the groundwork for new rules against AI tech deemed to be of ‘high risk.’
The outlines introduced by the Commission last year held back from introducing strict safeguards against facial recognition technologies, after an earlier leaked version of the paper floated the idea of a moratorium on facial recognition software.
The Commission instead opted to “launch an EU-wide debate on the use of remote biometric identification,” of which facial recognition technologies are a part.
Such a move was met with frustration by civil society groups, who had called for more stringent safeguards against certain technologies, including facial recognition applications, employed in public spaces.
Meanwhile, this week, digital rights group EDRi penned a letter to a group of EU Commissioners, calling for ‘regulatory limits on deployments of artificial intelligence that unduly restrict human rights.’ Of particular concern for EDRi is the use of biometric technologies, predictive policing applications, and the use of AI in border control activities.
For their part, however, the Commission has not ruled out a future ban on the use of facial recognition technology in Europe. Speaking to MEPs on the European Parliament’s Internal Market Committee in September last year, Kilian Gross of the Commission’s DG Connect said that all options were still on the table.
Responding to a question from Pirate MEP Marcel Kolaja on whether a potential ban is still in the offing, Gross said that “we will not exclude any option, we will look into all options and will carefully analyze existing legislation.”
More recently, Gross appeared in a panel held by EURACTIV in partnership with Swiss insurance firm, Zurich. He noted how the most preeminent risk emanating from AI applications concerned those related to ‘fundamental rights and safety.’
For S&D MEP Eva Kaili, meanwhile, safeguarding against such risks involves allowing an environment where trustworthy, ethical AI can flourish.
“We are trying to achieve a transformation that is ethical by design, where humans are not replaced but complemented,” she said during the event.
On this point, Rui Ferreira of Zurich was in accordance but added that the Commission’s future objectives for AI should also focus on the goal of harmonizing the bloc’s digital single market.
“The development of a future regulatory framework must enable the delivery of the EU Digital Single Market Initiative objectives,” Ferreira said.
“It must be adapted to the risks and investments in innovation while respecting societal values and norms”.
[Edited by Benjamin Fox]