Center for AI and Digital Policy

?? CAIDP Responds to Outcome of U.S. Convening of AI Safety Institutes The Center for AI and Digital Policy issued a statement to the International Network of AI Safety Institutes, warning that "AI safety science cannot advance in opacity" and "AI policy, including AI safety policy, should uphold democratic values in both the goals pursued and the processes established." Following the network's inaugural San Francisco meeting, CAIDP highlighted a "conspicuous lack of information about the agenda, objectives, speakers, and participants." The statement underscores that "in the rush to address existential risk there is the danger that the existing impact of AI on decisions in housing, credit, employment, education, and criminal justice will be ignored." We urged participating nations and institutes to ensure that "technical standards build on established AI governance frameworks and are rooted in human rights, democracy, and the rule of law." The CAIDP statement emphasizes that "it is critical to the AI safety agenda that AI risk is broadly understood to include fraud, bias, and other harms that could result from the deployment of insufficiently tested AI systems." CAIDP urged nations to sign the Council of Europe AI Treaty while also advocating for transparent processes that align with established frameworks, including the UNESCO Recommendation on AI Ethics and the EU AI Act. #AIsafety #AIfairness Council of Europe AI Safety Institute UNESCO Marc Rotenberg Merve Hickok Evelina Ayrapetyan Christabel R. Nidhi Sinha Jaya V.

?? CAIDP Advises US Senate on AI and Consumer Protection In a detailed statement to the U.S. Senate Committee on Commerce, Science, and Transportation Majority, the Center for AI and Digital Policy urged Senators to act on pending complaint before the Federal Trade Commission and to move forward AI legislation. "Over a year ago, CAIDP filed a detailed, formal complaint with the FTC about OpenAI, alleging that OpenAI had violated U.S. consumer protection law by releasing consumer products without sufficient safeguards. We described real risks to the public — giving bad investment advice or telling children how to lie to their parents. "In July 2023, both The New York Times and The Wall Street Journal reported that the FTC had launched the investigation the Center for AI and Digital Policy requested. "We ask that this Committee now urge the FTC to conclude the investigation of OpenAI, and issue an order. " ? We also ask the Committee to move forward bipartisan legislation that would safeguard consumers, including ?? S.3312-Artificial Intelligence Research, Innovation, and Accountability Act (ARIA) 2024 (Thune / Klobuchar) - requires transparency reports, risk management assessments, and enforcement against non-compliant high-risk applications ?? S. 4769 – the Validation and Evaluation for the Trustworthy Artificial Intelligence Act (Hickenlooper / ?? S. 4178- Future of Artificial Intelligence Innovation Act of 2024 (Cantwell, Young, Hickenlooper, Blackburn) ?? H.R. 5077 – the CREATE AI Act of 2024 (Heinrich, Young, Eshoo, Obernolte) CAIDP concluded, "We urge you- Chair Hickenlooper and Ranking Member Blackburn to work with Senate leader Schumer and McConnell to pass these bills during the lame-duck session." The Center for AI and Digital Policy statement to the Committee was included in the hearing record. Merve Hickok Marc Rotenberg Christabel R. Janhvi Patel #AIlegislation

?? ?? G20 Leaders Issue Declaration, Address AI The Leaders of the G20, met in Rio de Janeiro on 18-19 November 2024 to address major global challenges and crises and promote strong, sustainable, balanced, and inclusive growth. ???? The G20 Leaders made several key statements about Artificial Intelligence: ? "Transparency, with appropriate safeguards, and explainability regarding data, algorithms and content moderation that respects intellectual property rights and privacy, and data protection can be key for building healthy information ecosystems." (par. 29) ? "It is our endeavor to leverage AI for good and for all by solving challenges in a responsible, inclusive and human-centric manner, while protecting people's rights and safety." (par. 77) ? "To ensure safe, secure, and trustworthy AI development, deployment and use, the protection of human rights, transparency and explainability, fairness, accountability, regulation, safety, appropriate human oversight, ethics, biases, privacy, data protection and data governance must be addressed." (par. 77) ? "We reaffirm the G20 AI principles and the UNESCO Recommendation on Ethics of AI." (par. 77) ? "We recognize that technologies’ integration in the workplace is most successful when it incorporates the observations and feedback of workers and thus encourage enterprises to engage in social dialogue and other forms of consultation when integrating digital technologies at work." (par. 78) The G20 leaders welcomed the African Union as a full member of the G20. "Africa’s voice should be amplified in the G20 and all other international fora," stated the G20 Leaders. #aigovernace #g20 AfricanUnion CAIDP has provided extensive advice to the G20 regarding AI Policy over several years. Merve Hickok Nidhi Sinha Caroline Friedman Levy Monique Munarini April Yoder, PhD Marc Rotenberg

?? CAIDP Update 6.44 - AI Policy News (Nov. 18, 2024) ???? EU Launches Major Consultations on AI Act Implementation ???? ?EU Hits Meta With €798 Million Fine Over Facebook AI Practices ???? ???? Netherlands and Denmark Face Backlash Over Biased AI Systems ?? OECD Warns of AI Risks, Calls for Urgent Global Oversight ???? Canada Launches Federal AI Safety Institute ???? DHS Unveils Safety Guidelines for AI in Critical Infrastructure ??? ??? ??? ?? CAIDP Urges U.S. Senate to Ratify AI Treaty ??? ?? ?? ?? CAIDP Calls for Immediate Ban on Autonomous Weapons ??? ?? ???? CAIDP and Democracy Research Host AI Workshops in Turkey ?? New Book on AI Procurement by CAIDP President Merve Hickok ??? ?? ?? CAIDP Speaks at World Bank AI Governance Conference ?? CAIDP Spring 2025 AI Policy Clinic Applications – Deadline Approaching! European Union #netherlands #denmark #canada U.S. Department of Homeland Security OECD.AI The World Bank Christabel R. #AIsafety #AItreaty United Nations Human Rights

?? BREAKING - CAIDP Urges U.S. Senate to Ratify Council of Europe AI Treaty (Nov. 18, 2024) ??? ?? ??? The Center for AI and Digital Policy has sent a letter to the United States Senate Committee on Foreign Relations, urging the Committee to schedule a hearing for the ratification of the Council of Europe AI Treaty. The aim of the AI Treaty is to safeguard human rights, democracy, and the rule of law. CAIDP President Merve Hickok, who participated in the drafting and development of the Treaty, said "We urge the Committee to convene a hearing to move forward with the ratification of the Council of Europe AI Treaty. The longer the delay in ratification of the AI Treaty, the more difficult it will be to establish necessary safeguards that enable innovation and technological progress." ??? Ms. Hickok testified before the US Congress in March 2023, urging law makers to address the emerging challenges of AI. ???? In the letter to the Senate, CAIDP pointed out that the AI Treaty will advance the common goals set out by both the Biden and Trump administrations. "Now is the opportunity to advance these goals. The United States was instrumental in the development of the AI Treaty. The Treaty is aligned with the AI policies of successive U.S. administrations, American values, and American leadership in technological innovation. The United States and 35 other countries have signed the AI Treaty." ?? CAIDP wrote, "The crucial next step is for the Senate to ratify the AI Treaty. This must be a priority for the lame-duck session." #aigovernance #AItreaty Marc Rotenberg Christabel R.

It is competitive but it is so worth it! If you are interested in centering human rights and democratic values in AI policy and governance, applications are open!!!

?? Encode Justice Hosts "Empowering Voices Summit" - Nov 19 - Washington DC Our friends with Encode Justice are hosting an important policy event in Washington DC with Francesca Mani, who became an outspoken advocate after her high school classmates made deepfake NCII of her when she was 14. Senators Durbin, Cruz, and Klobuchar will be speaking alongside civil society experts like Mary Anne Franks, JD, DPhil, Yiota Souras, Omny Miranda Martone Stefan Turkheimer and representatives from the Kent School and Snap Inc. As Encode Justice policy director Adam Billen explains, "I care very deeply about all of the work we do at Encode, but this is the issue that makes me the most viscerally angry. We are spiraling towards a world where deepfake porn becomes a normal part of our students’ lives — already 15% of students know a victim in their school. The moment we are currently living in is the key point at which we have a chance to prevent that reality. I hope we will not miss it." We hope our friends and colleagues in Washington DC are able to attend this important event. ?? "Empowering Voices Summit: Protecting Students from Deepfake Exploitation" ?? Tuesday, November 19, 2024 ?? Senate Visitor Center 212-10 Sneha Revanur https://lnkd.in/e7w-Vm5b

?? Comments Requested on EU Draft General-Purpose AI Code of Practice ???? ?? This first draft of the Code addresses key considerations for providers of general-purpose AI models and for providers of general-purpose AI models with systemic risk, through four Working Groups working in close collaboration: WG1 - Transparency and copyright-related rules WG2 - Risk identification and assessment for systemic risk WG3 - Technical risk mitigation for systemic risk WG4 - Governance risk mitigation for systemic risk The Chairs and Vice-Chairs of the Working Groups said that the first draft is "a foundation for further refinement. Following an iterative process of internal discussions within the Working Groups and additional external input from stakeholders, Measures may be added, removed, or modified. "We invite stakeholders to review the document and provide feedback to help shape the final version of the Code, which will play a crucial role in guiding the future of general-purpose AI model development and deployment." ?? Deadline for comments: Thursday 28 November 12:00 CET https://lnkd.in/dAkwrQqy The Center for AI and Digital Policy Europe is one of the participating organizations drawing-up the first General-Purpose Code of Practice. CAIDP Europe applied to the Call for Expression of Interest and was considered eligible by the AI Office following the criteria outlined in the call. CAIDP Europe previously provided recommendations regarding procedures for developing the Code of Practice. The Center for AI and Digital Policy Europe "Draft Rules of Procedure for the GPAI Code of Practice Kick-off Plenary participants" (17 September 2024), Karine Caunes Giovanni De Gregorio ?i?dem Ak?n, LLM,MA Stamatios Anoustis Uche Anyamele, PhD Rose Boutboul Daniela Constantin Daniela Duta Monique Munarini https://lnkd.in/e5w--bGx #aigovernance #CodeOfPractice #EUAIA Ursula Pachl Selim Alan Merve Hickok

?? CAIDP Speaks at World Bank Conference on "Current Trends in AI Use and Governance" (Nov. 13, 2024) CAIDP Associate Director Christabel R. spoke this week at a conference organized by The World Bank on "AI in Practice: Current Trends in AI Use and Governance." The session was part of the World Bank Legal Vice Presidency’s Learning Week.?Many of the participants are lawyers, focusing on arrangements between the Bank and Recipients. Christabel outlined current AI governance frameworks, including the European Union AI Act and the Council of Europe treaty, as well as procurement standards?and the need for red lines for privacy-impacting and safety-impacting AI systems that fail to comply with minimum practices. In the last few years, CAIDP has spoken at several World Bank conferences, addressing lawyers, managers, and data protection officials about the emerging norms for the governance of AI. #aigovernance Marc Rotenberg #GDPR #AIA Conrad C. Daly Anne Joséphine Flanagan, MBA Adele Barzelay Jamie VanDodick

?? CAIDP / CAIDP Europe Welcomes the 'EU Commission Consultation on the Application of the Definition of an AI System and the Prohibited Practices Established in the Act - Encourage CSOs and AI Policy Experts to Express Their Views' ???? ?? The European AI Office is launching this multi-stakeholder consultation on the application of the definition of an AI system and the prohibited AI practices established in the AI Act. According to the the AI Office, this consultation is targeted to stakeholders of different categories, including providers and deployers of AI systems such as businesses, authorities, academia and research institutions, trade unions and other workers' representatives, civil society organisations, public supervisory authorities, and the general public. The Center for AI and Digital Policy Europe and the Center for AI and Digital Policy welcome this announcement. CAIDP has been actively engaged in the public discussion of the definition of AI systems for many years and has also routinely called for "red lines" on AI systems that violate human rights, are unsafe, or no longer under human control. A few recommendations for those who are providing comments ?? Elements of the Definition of an AI System (Article 3(1)) ? On the issue 'varying levels of autonomy,' (#2) we urge commentators to remind the AI Office that a zero (O) level of autonomy is still an AI System subject to the Act. ? On the issue of 'infers from the input it receives' (#5), we object to the use of the pronoun 'it' and remind the AI Office that AI systems do not have personality. ? We also ask the AI Office to clarify that 'rule-based' / deterministic / symbolic AI systems are included within the definition of an AI System ? AI systems that are safety-impacting or rights-impacting are not 'simple software' ?? Questions Relating to the Prohibitions (Article 5) ? Anticipating the rapid evolution of AI systems, we urge expansive interpretations of prohibited AI systems ? Anticipating the application of AI techniques to circumvent the goals of the AI Act [?? ?? ?? ], we urge the AI Office to "red team" the recommendations proposed. ? Recognizing the previous judgment of the Court of Justice of the European Union in Ligues des droits humains, C-817/19, we urge the AI Office to make clear that: - 'algorithms must function transparently and the result of their application must be traceable' - 'competent authorities must ensure that the person concerned is able to understand how those criteria and those programs work’ to determine whether a decision was unlawful or discriminatory - there must always be pre-determined criteria, subject to human review and determination, for any decision impacting fundamental rights #PublicVoiceOpportunity #AIgovernance Merve Hickok Christabel R. Marc Rotenberg Nidhi Sinha https://lnkd.in/eMheTjAg