Tertius News is an AI-native newsroom: an AI model reads the linked source articles below and extracts what each outlet reported, so you can compare their coverage side by side. How this works →
EU negotiators reach provisional deal on revised AI rules
EU member states and European Parliament lawmakers have reached a provisional agreement on changes to the AI Act, including delays to high-risk rules and a ban on unauthorized deepfake apps. The deal, part of the Commission's Digital Omnibus, aims to ease compliance but has drawn criticism from privacy advocates.
European Union member states and European Parliament lawmakers have reached a provisional deal on changes to the bloc's landmark artificial intelligence rules, according to Deutsche Welle. The agreement, announced early Thursday after late-night negotiations, is part of the European Commission's Digital Omnibus package aimed at simplifying digital regulations. However, The Hindu reported that talks had failed after 12 hours of negotiations and would resume next month, presenting a conflicting picture of the outcome.
Coverage Comparison
Deutsche Welle reported that the provisional deal includes delaying implementation of some AI obligations to prevent legal uncertainty, as stated in a European Parliament press release. The Hindu, on the other hand, said that EU countries and lawmakers failed to reach a deal and would reconvene next month. The Hindu's report also framed the changes as part of the Digital Omnibus, echoing the same policy backdrop but describing the negotiations as unsuccessful.
Key Claims
The provisional agreement postpones rules on high-risk AI systems, including those involving biometrics, critical infrastructure, education, employment, law enforcement, and border management, to December 2, 2027, from a prior deadline of August 2 this year, as reported by Deutsche Welle.
Machinery will be excluded from the AI Act, according to Deutsche Welle, with the understanding that it only needs to comply with sectoral safety rules it already faces.
A mandatory watermarking of AI-generated content will apply from December 2, as reported by Deutsche Welle.
Negotiators agreed to prohibit AI applications used to create unauthorized sexually explicit deepfakes, which could include images, video, or audio, according to Deutsche Welle.
The changes to the AI Act, which entered into force in August 2024 with staged enforcement starting this year, are part of the European Commission's Digital Omnibus, aimed at simplifying digital sector regulations to help businesses compete with U.S. and Asian rivals, as reported by The Hindu.
Perspectives
EU Council Presidency
Marilena Raouna, Deputy Minister for European Affairs of Cyprus, which holds the EU Council's rotating presidency, welcomed the deal, saying it "significantly supports our companies by reducing recurring administrative costs" and "ensures legal certainty and a smoother and more harmonized implementation of the rules across the Union."
Critics and Activists
Deutsche Welle noted that critics view the move as giving in to pressure from businesses. The Hindu reported that privacy activists and civil rights groups have criticized the proposed changes as "caving to Big Tech." The Hindu's report did not include direct quotes from these critics.
How each outlet told it
Deutsche Welle
Framing: EU reaches deal on watered-down AI rules — Informative
Facts Included:
EU member nations and European Parliament lawmakers reached a provision deal on AI rules
The deal includes delaying implementation of some obligations for AI systems
Rules on high-risk AI systems will be postponed to December 2, 2027
Machinery will be excluded from the AI act
A mandatory watermarking of AI-generated content will apply from December 2
A ban on AI applications used to create unauthorized sexually explicit deepfakes was agreed upon
AI-extracted; can misattribute a claim — see Methodology.
Each row is one claim, attributed to the outlet whose wording states it most clearly. Confidence rates how directly the source text states the claim — explicit and unhedged rates high; hedged, pieced-together, or internally inconsistent statements rate lower. It does not measure whether the claim is true. Status counts the distinct outlets we found asserting it — so a single-source claim can still show high confidence, and a multi-source claim can show medium. Every one of those outlets is named beside the status, so you can check the count against the list. For claims extracted before we began storing that list, the row says so: it names the outlet the claim is quoted from and states that we have not recorded which outlets backed it. Outlets wrote at different times, so a figure that evolves — a casualty count, for example — can legitimately differ between rows; check the "as of" time next to each claim's source.
Claim
Confidence
Status
ClaimEU member nations and European Parliament lawmakers reached a provision deal on AI rules. The deal is part of the European Commission's digital omnibus package.