EU lawmakers agree to ban facial recognition use in public and predictive policing tools, and set transparency rules on generative AI; details must be finalized
The significance is that public-sector surveillance tools and generative-AI transparency are being addressed within the same legislative track, though the final terms remain unsettled.
First-order effects
Police and other public authorities face a prospective prohibition on using facial recognition in public spaces and predictive-policing tools, subject to the final text.
Generative-AI providers face prospective transparency obligations, while lawmakers and the EU Commission must resolve the provisions' operational details.
Second-order effects
Vendors selling biometric surveillance or predictive-policing systems into Europe will need to reassess whether those products can be offered to public-sector customers under the eventual rules.
AI providers will have an incentive to prepare disclosure processes early, since the same framework is pairing generative-AI obligations with restrictions on sensitive public-safety uses.
Third-order effects
If finalized in this form, the EU framework would reinforce a governance model that distinguishes between AI uses considered unacceptable and those allowed only with obligations, rather than treating AI as a single regulated category.
The unresolved details will determine whether compliance becomes a durable market-access requirement or leaves meaningful room for national interpretation and exemptions.
The trend: This is part of a broader shift toward regulating AI by use case, with public surveillance receiving the strongest constraints and general-purpose generative systems facing transparency duties.
All over in a heartbeat, all compromises on the EU #AIAct passed. That means full bans on facial recognition in public, predictive policing, fundamental rights impact assessments, transparency for general purpose AI, and rights and redress for affected persons https://twitter.com…
EU lawmakers approved a new version of its upcoming AI Act with strong prohibitions on mass biometric surveillance, predictive policing and much more. Digital rights campaigners tell me it's “overwhelmingly good news.” Details here: https://www.theverge.com/...
Am I reading this right? The AI Act won't apply to open source systems (unless integrated into high-risk systems) ... *except* open source foundation models? Uhhh...🤔 https://twitter.com/...
We need EU-wide rules for safe use of AI... Facial recognition is a slippery slope. History shows the boundaries are quickly moved. Let's regulate smartly but firmly... before others 🇨🇳 🇺🇸 move far beyond what fits our 🇪🇺 values ! https://www.theguardian.com/ ...
“rights experts have argued that surveillance does not equal safety, and ...these uses of AI ... redirect policing and surveillance into already marginalized and over-policed communities.” Myself + @djleufer spoke to @jjvincent @Verge on the #AIAct https://www.theverge.com/...
BREAKING: The European Parliament leading committees adopted the #AI Act with a broad majority. The EPP's attempt against the ban on remote biometric identification was not successful. Next step: plenary vote on 14 June (TBC).
This is going to make @FiveThirtyEight or @TheEconomist style predictive journalism around elections much more difficult, given the potential of those simulations to influence voters. https://twitter.com/...
High risk AI now includes: “AI systems to influence voters in political campaigns and in recommender systems used by social media platforms (with more than 45 million users under the Digital Services Act)” https://www.europarl.europa.eu/ ...
Artificial Intelligence Act: MEPs in @EP_SingleMarket and @EP_Justice endorse new transparency and risk-management rules aiming to ensure human-centric and ethical AI development. Next step? June plenary vote to endorse EP mandate. Press release⤵️ https://eptwitter.eu/qCpt
#AI Act: a step closer to the first rules on Artificial Intelligence To ensure a human-centric and ethical development of Artificial Intelligence (AI) in Europe, MEPs endorsed new transparency and risk-management rules for AI systems. Details ⬇️ https://www.europarl.europa.eu/ ..…
#AI Act: Following the @EP_SingleMarket and @EP_Justice committees' adoption of the file, @eaTechEU provides an overview of the main changes the European Parliament introduced to the text. https://www.euractiv.com/...