Center for AI and Digital Policy Europe

🔥 TODAY, August 1st, THE #EUAIAct ENTERS INTO FORCE https://lnkd.in/egkvkTHP Karine Caunes, CAIDP Europe's Executive Director stated: "We created the European branch of the Center for AI and Digital Policy with a clear goal in mind: to ensure a rights-based implementation of the EU AI Act. The time is NOW." ⏰ The clock is ticking. The EU AI Act will come gradually into application. ⭐ AI Governance 👉 Implementation with the common good of all in mind & not left to self-regulation whether through standardisation for high-risk AI systems or code of practice for GPAI models 👉 The EU AI Office and their national counterparts should ensure an inclusive implementation process + also Re. the AI Pact (https://lnkd.in/ddnwsBuF) 🛎 See also AI & risks literacy for employees of all organisations using AI, deadline February 2, 2025 See on the 1st GPAI Code of Practice drafting https://lnkd.in/eYHRhXzS 🛎 By November 2, 2024: Member States shall identify the authorities responsible for fundamental rights protection Article 77(2) EU AI Act 🛎 By August 2, 2025: Member States shall designate National Market Surveillance Authorities. They should be COMPLETELY independence See our study https://lnkd.in/eaMEGp5u See CSOs collective call https://lnkd.in/e26TZaBK See EDPB statement: https://lnkd.in/dGSf7Egy ⭐ Substantive rules 👉 Compliance with EU AI Act is NOT enough for practices to be legal. They need to: ✅ comply with existing European & national human rights law. See Council of Europe's European Convention on Human Rights + recently drafted CoE Convention on AI See EU Charter of Fundamental Rights & EU legislation (consumer protection, employment, non discrimination...) https://lnkd.in/emWD648j ✅ Comply with GDPR + other relevant rules of DSA, DMA... 🛎 By February 2, 2025: provisions on AI PROHIBITED practices apply: + subliminal or deceptive techniques + exploitation of vulnerabilities + social credit scoring + predictive policing + emotion recognition in the workplace & education + facial recognition databases + biometric categorisation to infer protected characteristics + law enforcement use of real-time biometric identification in public space CAIDP Europe will ensure these provisions are not interpreted restrictively and converted into high risks practices.

🌐 Swiss AI Regulation in Focus 🌐 The 🇨🇭 Federal Council is actively reviewing potential approaches to regulating artificial intelligence (AI) until the end of 2024. Following this assessment, measures will be taken to ensure responsible and fair use of AI technologies. At the Center for AI and Digital Policy Europe, we recognise the importance of these efforts. We support the push for robust regulations that prioritise human rights and protect against the risks of AI-driven discrimination. We also encourage our network to join us in supporting initiatives like the one by AlgorithmWatch CH for non-discriminatory AI. By working together, we can shape a future where AI benefits everyone, without compromising fairness and equality. “At CAIDP Europe, we believe that the benefits of AI should be accessible to all, without compromising fundamental rights or deepening societal inequalities. This aligns closely with the goals of the "Artificial Intelligence Without Discrimination" campaign by AlgorithmWatch Switzerland” said Giuliano Borter, CAIDP Europe Senior Policy Officer and Administrative Director. To know more and support this initiative, please visit https://lnkd.in/e6W77uMF #AIPolicy #EthicalAI #AIRegulation #HumanRights #SwissAINews #AIwithoutDiscrimination

🔥 🔥 🔥 European AI office x Trustworthy General Purpose AI models under #EUAIAct. Today the European AI Office is launching two key initiatives regarding the drafting of the first EU GPAI Code of Practice. Karine Caunes, CAIDP Europe Executive Director said, "CAIDP Europe is committed to contribute in both initiatives to ensure that the EU AI Act is implemented to fulfill its goal to preserve the rights of all and serve the good of all." CAIDP Europe calls on civil society organisations, academics, independent experts and concerned citizens to contribute as well. Firstly, a public consultation: https://lnkd.in/efG8zVjD WHAT? 👉 The topics covered by the Code of Practice detailing rules for general-purpose AI models providers 👉 A questionnaire with three sections: 1. General-purpose AI models: transparency and copyright-related provisions; 2. General-purpose AI models with systemic risk: risk taxonomy, assessment and mitigation; 3. Reviewing and monitoring the Codes of Practice for general-purpose AI models. Bonus section: carte blanche / any other relevant aspect to share. FOR WHOM? All interested parties. E.g. civil society organisations, academia, independent experts, industry representatives, rightsholders, and public authorities. WHEN? Opening: 30 July 2024 Closing: 10 September 2024 WHY? To inform the initial draft of the Code of Practice Secondly, a call for expression of interest: https://lnkd.in/ezSsv52t WHAT? To participate in the drawing-up of the GPAI Code of Practice. The Code of Practice will detail the AI Act rules for providers of GPAI models and GPAI models with systemic risks. Providers should be able to rely on the Code of Practice to demonstrate compliance. FOR WHOM? Eligible general-purpose AI model providers, downstream providers and other industry organisations, other stakeholder organisations such as civil society organisations or rightsholders organisations, as well as academia and other independent experts. WHEN? 👉 Express your interest by 25 August 2024, 18:00 CET. 👉 Iterative drafting process by April 2025. 👉 Code of practice due to apply by August 2nd, 2025. After publication of the Code, the AI Office and the AI Board will assess its adequacy and publish this assessment. The Commission may decide to approve the Code of Practice and give it a general validity within the Union by means of an implementing act. If by August 2nd, 2025, the Code of Practice cannot be finalised or the AI Office deems it not adequate, the Commission will provide common rules for the implementation of the relevant obligations. WHY? Art. 56 EU AI Act: "The AI Office shall encourage and facilitate the drawing up of codes of practice at Union level in order to contribute to the proper application of this Regulation, taking into account international approaches."

🚨 National Court Decision on Facebook: When Competition Law Safeguards Digital Rights🚨 Recently, the Düsseldorf Regional Court made a noteworthy decision regarding Facebook (Meta) and its treatment of users. The court ruled that Meta violated competition law by blocking a user's Facebook account without proper justification. This decision highlights important issues about how big tech companies should operate and protect users' rights. Jurisdiction: Despite Meta’s terms of use suggesting disputes be handled in Ireland, the court ruled it had jurisdiction due to the competition law aspect, allowing German users to challenge unfair practices locally. Abuse of Market Dominance: The court found Meta abused its dominant position by blocking the account without clear and immediate justification, distorting competition and impacting the association's ability to communicate. Need for Transparency: Emphasizing EU regulations like the P2B Regulation and the DSA, the court stressed that platforms must be transparent and fair, providing clear reasons for actions taken against users. Balancing Interests: The court considered factors like the imbalance of power, the social significance of Facebook, and the plaintiff's freedom of expression, finding Meta’s actions unjustified. Policy Implications: 🚩 Transparency and Fairness: Companies must provide clear and specific reasons for blocking accounts. 🚩 User Rights: Users can challenge unfair practices in their own country’s courts. 🚩 Complementary Laws: German competition law and EU regulations like P2B and DSA work together to protect users. Conclusion: This decision is a major step towards ensuring fairness in the digital world, setting an important precedent for competition law and digital rights. Read more here https://lnkd.in/exyUT5y6 #DigitalRights #CompetitionLaw #Facebook #Transparency #UserRights #CAIDPEurope

📢 European Parliament Sets Up Working Group to Oversee Implementation of the AI Act According to both Euractiv and MLex, the Internal Market and Consumer Protection (IMCO) Committee and the Civil Liberties, Justice and Home Affairs (LIBE) Committee will set up a Working Group to ensure effective implementation of the AI Act. The European Commission plays a central role in the implementation of the AI Act. The Parliament is seeking to ensure the Commission fulfills is its responsibilities. MLEX - "The initiative follows the footprints of similar working groups that were set up to monitor the enforcement of the Digital Services Act and the Digital Markets Act and maintain pressure on the European Commission." Euractiv - "Members of the European Parliament have previously expressed concerns over the lack of transparency around the AI Office’s staffing process, as well as the involvement of civil society in parts of the implementation process." Members of the Center for AI and Digital Policy met with members of the IMCO and LIBE committees during the development of the AI Act, provided detailed comments on the legislation, and then published a report for the European AI & Society Fund "Making the AI Act work: How civil society can ensure Europe’s new regulation serves people & society" (September 2023) - https://lnkd.in/gaZDTUas There are several upcoming deadlines for the AI Act: 6-months after adoption (Feb 2025) ➡ Ban on high-risk AI systems (subliminal techniques, exploitive technqiues, web scraping of facial images) 🔥 🔥 🔥 ➡ AI literacy obligations published 12-months (August 2025) ➡ Obligations for General Purpose AI Systems established ➡ Serious incident reporting guidance published ➡ Enforcement provisions established 18-months (Feb 2026) ➡ Guidance for High-risk AI Systems published European Parliament European Commission #IMCO #LIBE #aigovernance #aiact #euaiact Center for AI and Digital Policy Europe https://lnkd.in/emQ65r54

🔥 🔥 🔥 34 CSOs, including CAIDP Europe, call on the European Commission and Member States to fulfill their obligation to ensure the independence of national market surveillance authorities (NMSA) under the EU AI Act. Karine Caunes, CAIDP Europe's Executive Director stated: "We welcome the Commission's and the Member States' efforts to set in place swiftly an AI governance framework. However, effective enforcement requires ensuring the COMPLETE independence of NMSA in compliance with the EU AI Act" ✅ CAIDP Europe is concerned that some States are designating bodies that are deprived of full independence from State authorities as NMSA ✅ NMSA's independence from both States & market actors is necessary to ensure the effective protection of fundamental rights. Fundamental rights violations can occur from either sources. ✅ NMSA COMPLETE independence is a LEGAL OBLIGATION imposed by the EU AI Act and already clarified by the European Court of Justice 👉 For more details see CAIDP Europe's analysis conducted under our strategic priority "Consolidating a human centric governance of AI" available here: https://lnkd.in/dNQBmZzT Among key issues: 🔸 The European Court of Justice has repetitively held: "OPERATIONAL independence is NOT sufficient in itself to protect supervisory authorities from all external influence (...) the mere risk that the State scrutinising authorities could exercise a political influence over the decisions of the supervisory authorities is enough to hinder the latter in the independent performance of their tasks" 🔸 The European Union Agency for Fundamental Rights, in its recent report on the data protection authorities' experience, shows that LACK OF RESOURCES undermines effective enforcement. FRA's report available here: https://fra.europa.eu/en 👉 The Commission and Member States' failure to respect their obligations under the EU AI Act would inevitably lead to new cases brought before the European Court of Justice with very little doubt regarding the outcome. ✅ Failure to respect the independence of NMSA could have a negative spill-over effect on the independence of Digital Services Coordinators under the DSA and Data Protection Authorities under the GDPR 👉 Some states are designating as NMSA data protection authorities or bodies named as Digital Services Coordinators. Two issues: 1/ undermining these supervisory authorities' independence 2/ asymetry in effective enforcement of the EU AI Act across Europe ✅ Only through collective intelligence and action can we ensure the human centric governance of AI and enforcement of the EU AI Act CSOs' collective call available here: https://lnkd.in/gB83u_av Center for AI and Digital Policy Rose Boutboul Giuliano Borter Selin Ozbek Cittone Uche Anyamele, PhD Lina Chtioui Marc Rotenberg Thierry Boulangé Frederico Oliveira da Silva Daniel Leufer Alessandra Bormioli

Join us tomorrow for a great discussion on Navigating AI Futures: A Comparative Look at International AI Rules together with OECD Julia M. Schmidt, IAPP Joe Jones & Digital Policy Alert Johannes Fritz. Spoiler alert: at the Center for AI and Digital Policy our key focus and leitmotiv is how to ensure human centric AI governance through international convergence nourished by a bottom up approach in order to tackle pressing challenges AI poses to human rights, democracy and the rule of law 📅 Date and Time: 20 June, 15:00 CET See details below 👇

CAIDP Europe executive Director Karine Caunes' participation in a prestigious panel on 'AI regulation in Europe' today in Athens was an opportunity to restate the absolute necessity to develop an integrated approach to all relevant rules at Council of Europe, EU and national level in order to ensure a rights-based approach to AI regulation. Dr. Karine Caunes stated, "a rights-based approach to AI regulation is the trademark of the "European way". This is necessary to protect human rights, democracy and the rule of law. This is what will allow us to develop a truly innovative AI path and this is what will protect us against unfair competition from Big Tech developing and putting on the market half-baked AI systems."

CAIDP Europe congratulates the Dutch Government and supervisory authorities for their transparency Re. shaping AI governance in the Netherlands. Some key takeaways from the second advice of the Autoriteit Persoonsgegevens (Dutch DPA) and the Rijksinspectie Digitale Infrastructuur to the Dutch Government on the EU AI Act supervisory structure in the Netherlands: ✅ Are involved not only ministries in charge of economic affairs or digitalisation but the Justice ministry as well. 👉 good practice taking into consideration that the EU AI Act is about managing risks and harms to fundamental rights 👉 these ministries do cover consumer protection ❓ Why are the ministries of social affairs and employment or the ministry of education not involved? Employment and education are key fields covered by the EU AI Act. ✅ Involvement of 22 supervisory agencies 👉 coordination among supervisory authorities is key to effective application & enforcement of the EU AI Act ✅ Proposed market surveillance authority in most fields: the AI Unit of the Dutch DPA. 👉 the Dutch DPA is an INDEPENDENT supervisory authority GDPR on independence is more protective than EU AI Act 👉 Can build on the expertise already acquired by the Dutch DPA (automated decision-making / algorithmic transparency under GDPR) and ensure direct coordination within the Dutch DPA. Moving forward, we might see an integration of the GDPR data protection impact assessments within the AI Act risk management matrix ⚠ The Dutch DPA.AI and other supervisory authorities involved should get the necessary human resources and financial means to carry their mission effectively ⚠ The Dutch DPA.AI should be the designated Dutch rep. within the European AI Board 👉 Effective enforcement of the AI Act requires for the same supervisory authority to ensure external coordination with other EU member states' market surveillance authorities and internal coordination with other Dutch supervisory authorities ⚠ full transparency requires public consultation on the Dutch AI governance framework For more news, check out https://www.caidp.eu/ Giuliano Borter Center for AI and Digital Policy Marc Rotenberg Rose Boutboul Natali Helberger Yelyzaveta (Lisa) Markova Jan-Jaap Oerlemans Stefan Roolvink M. Konrad Borowicz Gizem Yardımcı Alexander Laufer Zelal B. Iris Muis Esther Keymolen Brend P. Jurriaan Parie Sarah E. Gerald Hopster Emaediong Akpan Eleni Kosta Eleni Metanoia Evelien Brouwer Elena Kukovica Oskar J. Gstrein Mirko Tobias Schäfer Cosimo Monda Gijs van Dijck Panos DelimatsisEduard Fosch Villaronga Catelijne Muller Frederik Zuiderveen Borgesius Sven Stevenson Roel Dobbe Jef Ausloos Sofia Ranchordas

Venice Privacy Symposium - Remarks on AI Policy Convergence and Divergence Here are a few Speaker's Notes: - To those in the data protection field, policy convergence/divergence is a familiar theme. - In a global, network economy convergence should be the norm, with increasing regulation (=better safeguards) as challenges are better understood and political institutions mature - The Brussels Effect (and perhaps the Strasbourg Effect) helps explain increasing convergence - Important for the EU and US to align behind Democratic Values - Harmonized rules (= convergence) should be understood as "a floor and not a ceiling." - Governments should remain free to develop innovative legislation as new challenges emerge. The response to generative AI is particularly relevant. - The AI safety agenda should not displace the AI fairness agenda. Safety is critical but rights are fundamental. - The focus now should be implementation and enforcement of the emerging norms for the governance of AI As for the Center for AI and Digital Policy and the Center for AI and Digital Policy Europe - We provide an extensive review of AI policies and practices in our annual AI and Democratic Values Index - We support the effective implementation of the European Union AI Act - We will promote ratification and implementation of the Council of Europe AI Treaty(open for signature 5 September 2024) - We continue to endorse the "Universal Guidelines for AI"as a global baseline for #aigovernance ➡ We also recommend that the United Nations establish a scientific panel on AI, similar to the IPCC, and a Special Rapporteur for AI and Human Rights United Nations Human Rights IPCC