EU set to allow draconian use of facial recognition tech, say lawmakers

[Stone]

EU set to allow draconian use of facial recognition tech, say lawmakers
Late tweaks to the EU’s artificial intelligence law have caused uproar over loopholes for biometric identification.

JANUARY 16, 2024
BRUSSELS (Politico.eu) — Last-minute tweaks to the European Union’s Artificial Intelligence Act will allow law enforcement to use facial recognition technology on recorded video footage without a judge’s approval — going further than what was agreed by the three EU institutions, according to European lawmaker Svenja Hahn.

The German member of the European Parliament said the final text of the bloc’s new rules on artificial intelligence, obtained by POLITICO, was “an attack on civil rights” and could enable “irresponsible and disproportionate use of biometric identification technology, as we otherwise only know from authoritarian states such as China.”

The wording also made it to the full legal text, which the Spanish Council presidency put together on December 22. The current presidency of the EU Council, held by Belgium, is working with Parliament to finalize bits of interpretative text known as recitals.

In early December, the EU agreed on a pioneering artificial intelligence rulebook — the world's first — to slap wide-ranging binding rules on the use of the burgeoning technology. But Hahn, a member of the liberal Renew group, said the final wording of the text introduced a loophole for the use of facial recognition technology, which was not in the original agreement.

In a statement to POLITICO, Hahn outlined issues with final language on so-called post facial recognition, where the technology is used on pre-existing footage, distinguishing it from real-time scanning of public spaces with AI-augmented cameras, the use of which would be largely outlawed under the AI Act.

The Spanish presidency of the EU Council — representing member countries' governments — and aides from Parliament agreed the rules on post facial recognition on December 22, two weeks after the two institutions and the European Commission reached a common position on the AI rulebook as a whole.

Hahn argued that the text breached that December 8 agreement. “The oral agreement had foreseen the use of post [facial recognition] only for very serious crimes, under very strict conditions, such as a prior judicial reservation. Little of this remained,” Hahn said.

The German parliamentarian said the rulebook's final text would allow police forces to use post facial recognition after the say-so of an administrative authority, rather than a judge's decision. She also lamented that the technology would be allowed to identify suspects for all types of crimes, regardless of how severe these crimes are. “The most trivial misdemeanors could be prosecuted using facial recognition,” she said.

Hahn's concerns were echoed by European Parliament member Patrick Breyer, a member of the left-leaning German Pirate Party and self-proclaimed "digital freedom fighter," who said in a statement Tuesday that "it appears the EU intends to compete with China not only technologically but also in terms of high-tech repression."

The kerfuffle over facial recognition underlines how European Parliament and Council officials left the 36-hour final negotiation round on December 8 with different understandings of what had been agreed to regarding post facial recognition. That happened despite a version of the text being shown on the negotiating room’s screen, according to a diplomat familiar with the matter, granted anonymity to speak about confidential negotiations. The Spanish presidency of the EU Council did not respond to a request for comment.


Clarion call

“We had heard from both the Commission and the Parliament that post [facial recognition] would be subject to strict safeguards and only allowed in some narrow, exceptional circumstances,” said Daniel Leufer, a senior policy fellow at digital rights NGO Access Now. “The latest text that we’ve seen makes a mockery of those announcements.”

Ella Jakubowska, a senior policy adviser at European Digital Rights, said that the text opened “huge loopholes which allow the wide use of post biometric identification.”

But others, including Parliament's leading negotiators of the deal, defended the final text.

Dragoș Tudorache, one of two lawmakers who steered the AI Act through Parliament — and member of the same political group as Hahn — said the language on facial recognition was "supported by a majority" of leading parliament members and that it "reflects the political agreement reached" in December.

EU governments are slated to receive the AI Act’s final text on January 24, with the aim of green-lighting it on February 2. After that, the Parliament will have to pass the law with a plenary vote. Parliamentary committees, political groups or alliances of over 40 lawmakers will be able to propose amendments — which, if approved, will need further legislative work in Council and Parliament.