Center for AI and Digital Policy

📢 CAIDP California Speaks at Civil Rights Council about Automated Decisionmaking and Employment CAIDP Research Fellow Evelina Ayrapetyan testified at the California Civil Rights Council Public Hearing on July 18, 2024 regarding the Council’s proposed modifications to employment regulations for automated-decision systems. Ayrapetyan explained, "AI is quickly transforming the employment landscape and while it has great potential, without proper guardrails in place, the use of AI systems in employment can amplify existing biases and negatively impact historically marginalized communities." She made several recommendations. 1️⃣ Require public disclosure: 1) that a system is in use, 2) methods for opt in/out, 3) explanation of the system’s logic regarding the candidate or employee, and 4) results of the independent impact assessment algorithmic decision systems 2️⃣ Explicitly Require Human oversight and individualized assessments in final employment decisions, even when a job applicant/employee opts-in to algorithmic decision-making 3️⃣ Require pre-deployment independent algorithmic system impact assessments as a pre-condition to deploying AI systems or algorithmic tools in employment 4️⃣ Expand the definition of “Algorithmic-Decision System” and consider including the term ‘use of an algorithmic employment tool’ She concluded, "We applaud the Civic Rights Council for ensuring California employers deploy AI systems ethically and responsibly while respecting human rights and the rule of law, and protecting people in vulnerable situations who are most likely to be harmed by algorithmic decision-making." Evelina Ayrapetyan and several other members of the Center for AI and Digital Policy Research Group have launched a new California-based affiliate for CAIDP. The affiliate will represent CAIDP in state-level AI policy issues. The California legislature is currently considering several AI bills, including discrimination in housing and health care services, AI and employment, protections for creative artists, privacy and surveillance, and standards for foundational models. #aigovernance #california #employment Evelina Ayrapetyan Nidhi Sinha Jaya V. Christabel R. Merve Hickok Marc Rotenberg

📢 GPAI Publishes Report on Algorithmic Transparency in the Public Sector The report from the GPAI reviews algorithmic transparency instruments in the public sector and focuses on repositories or registers of public algorithms. The project's objective is to study algorithmic transparency in the public sector with an emphasis on assessing transparency instruments that enable governments to comply with algorithmic transparency principles, standards, and rules. The GPAI report explains that "algorithmic transparency arises within the broader context of public interest regulation. The principle derives from the democratic right to know and access information." 🔥 GPAI - "Algorithmic transparency is a means for fulfilling fundamental rights enshrined in public interest regulation. Applied to the public sector, for example, information on how state services are provided enables the population to access health and education rights. Moreover, information about how the state makes certain decisions affecting people's llives and liberties is indispensable to protecting the right to due process." 🔥 GPAI - "transparency in the public sector is one of the pillars of Open Government initiatives that governments worldwide have pledged to promote. . . . algorithmic transparency has become central to the new generations of Open Government initiatives." 🔥 GPAI - "algorithmic transparency enables citizen oversight over governmental activities and decisions associated with the adoption and implementation of ADM systems. For example, accessing meaningful information may allow civil society organizations to assess whether the use of ADM system complies with the law." The Center for AI and Digital Policy welcomes the GPAI report on Algorithmic Transparency ➡ Algorithmic transparency is one of the key metrics in our annual evaluation of national AI policies and practices in the CAIDP "AI and Democratic Values Index" ➡ The GPAI Report responds to the urgent need to move from principles to action to promote algorithmic transparency and accountability ➡ The GPAI Report builds on well-established principles of citizen access to information about government-decisionmaking ➡ The Center for AI and Digital Policy has previously advised international organizations to promote algorithmic transparency as part of AI governance. In 2021 and 2023, we asked the #G20 nations "to promote fairness, accountability, and transparency for all AI systems, particularly for public services. G20 leaders should adopt new laws to ensure algorithmic transparency and to limit algorithmic bias so that unfair treatment is not embedded in automated systems.” CAIDP President Merve Hickok has written extensively about the need to promote accountability of AI systems in the public sector. Juan David Gutiérrez Rodríguez Alison Gillwald CEIMIA #aigovernance OECD.AI Daniela Constantin Nayyara Rahman

📢 Council of Europe Publishes Overview of AI Treaty 📘 The Council of Europe has published an important overview of the "Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law," the first international legally binding instrument for AI. ➡ The AI Treaty "aims to ensure that activities within the lifecycle of artificial intelligence systems are fully consistent with human rights, democracy and the rule of law, while being conducive to technological progress and innovation." ➡ Activities within the lifecycle of AI systems must comply with the following fundamental principles: ► Human dignity and individual autonomy ► Equality and non-discrimination ► Respect for privacy and personal data protection ► Transparency and oversight ► Accountability and responsibility ► Reliability ► Safe innovation 🔥 There is the possibility for the authorities to ban or establish moratoria on certain application of AI systems (“red lines”). 🔥 The Framework Convention covers the use of AI systems by public authorities – including private actors acting on their behalf – and private actors. ➡ The Convention offers Parties two modalities to comply with its principles and obligations when regulating the private sector: Parties may opt to be directly obliged by the relevant Convention provisions or, as an alternative, take other measures to comply with the treaty’s provisions while fully respecting their international obligations regarding human rights, democracy and the rule of law. ➡ Parties to the Framework Convention are not required to apply the provisions of the treaty to activities related to the protection of their national security interests but must ensure that such activities respect international law and democratic institutions and processes. The Framework Convention does not apply to national defence matters nor to research and development activities, except when the testing of AI systems may have the potential to interfere with human rights, democracy, or the rule of law. The Center for AI and Digital Policy worked closely with civil society organizations, academic experts, member delegations, and the secretariat on the development of the AI Treaty. We look forward to adoption and implementation. The treaty will be open for signature on September 5, 2024. #aigovernance #aitreaty Merve Hickok Karine Caunes Daniela Constantin Marc Rotenberg Christabel R. Francesca Fanucci Center for AI and Digital Policy Europe European Center for Not-for-Profit Law Stichting

🌪 White House Announces $100 m Funding to Advance Public Interest Technology (July 16, 2024) From the White House Office of Science and Technology Policy: "The Biden-Harris Administration is committed to advancing technology that protects our safety, security, democratic values, and human rights. In his Executive Order on the Safe, Trustworthy, and Responsible Development and Use of AI, President Biden instructed the United States government to pull every lever to attract and hire highly skilled talent in AI and critical and emerging technologies." ➡ The National Science Foundation (NSF) will provide at least $48 million to advance research, implementation, and learning opportunities. ➡ The Department of Defense, with support from the Office of Management and Budget and OSTP, will launch the Trusted Advisors Pilot this year. ➡ The Ford Foundation is dedicating more than $20 million to enhance the field of public interest technology. ➡ The Siegel Family Endowment will invest $20 million in the public interest technology ecosystem over the next three years. ➡ The Mitchell Kapor Foundation is committing nearly $1.5 million to promote responsible, equitable, and ethical AI systems for the public interest. ➡ The University of Michigan Science, Technology, and Public Policy Program’s Community Partnerships Initiative will develop a collaborative AI innovation process with social justice organizations and local governments in southeast Michigan https://lnkd.in/gH6Fgckz

📢 CAIDP Update 6.28 - AI Policy News (July 15, 2024) ✝ 🕌 ☸ 🕉 ✡ World Religions Unite in Hiroshima to Sign AI Ethics Pledge 🇺🇸 FTC Outlines AI Oversight as CAIDP Urges Action on OpenAI Investigation 🇪🇺 EU AI Act Published: Compliance Countdown Begins for Companies NATO Unveils New AI Strategy in Tech Transformation Push 🇧🇷 Brazil Maintains Ban on Meta's AI Data Collection 🇩🇪 German Bundesrat Proposes Law to Criminalize Deepfakes 🗣 🏛 CAIDP Report Urges Congress to Act on FTC's Stalled OpenAI Investigation 🗣 🇬🇧 CAIDP President Addresses AI Challenges at Data for Policy Conference 🗣 🏛 CAIDP Urges Senate Action on AI Privacy Protections #aigovernance #euaia #brazil #germany Merve Hickok Federal Trade Commission NATO

What a privilege to meet alumni of Center for AI and Digital Policy in conferences across the globe. Simone Maria Parazzoli and Joseph A. shining through with their research on algorithmic accountability and data custodians 👏🏽 And finally meeting in-person with our CAIDP Fellow and incoming Teaching Fellow Selin Ozbek Cittone 🫶🏽 Feeling grateful for this community Marc Rotenberg Karine Caunes Grace S. Thomson Varsha Sewlal Tamiko Eto Selim Alan Also thank you Data for Policy CIC team for great conversations - Zeynep Engin, PhD Jon Crowcroft Stefaan Verhulst, PhD Mark Kennedy Peter Ford

📢 CAIDP Delivers Statement to US Senate on AI and Privacy In advance of a Senate hearing on "The Need to Protect Americans’ Privacy and the AI Accelerant," the Center for AI and Digital Policy submitted a statement with several recommendations. ➡ CAIDP expressed support for baseline, federal privacy legislation, and particularly provisions in the original American Privacy Rights Act that would ensure transparency and accountability for AI-based decision-making. ➡ CAIDP emphasized the correct understanding of Privacy Enhancing Technologies are techniques that "minimize or eliminate the collection of personal data." These techniques safeguard privacy, minimize the risks of data breaches, and begin to address the environmental impact of large AI models. PETs, properly understood, are the "anti-accelerant." ➡ CAIDP called attention to public survey data that shows broad public support - across regions and political parties - for government action to regulate AI ➡ CAIDP also called attention to its new report - "ChatGPT and the Federal Trade Commission: Still No Guardrails." In the report, CAIDP traced the history of its 2023 complaint to the Federal Trade Commission about OpenAI, the early reports of the investigation in The New York Times and The Wall Street Journal, and then the FTC's delay in issuing a final order. #aigovernance U.S. Senate Committee on Commerce, Science, and Transportation Majority Merve Hickok Marc Rotenberg Christabel R. Kyler Zhou

📢 CAIDP Releases "Still No Guardrails" Report - Urges Congress to Press FTC Commissioners on OpenAI Investigation 🔥 🔥 🔥 🗣 Prior to a Congressional hearing on the Federal Trade Commission, the Center for AI and Digital Policy released a new report - "Still No Guardrails" - to call attention to the FTC’s failure to act on the complaint CAIDP filed with the FTC about OpenAI in March 2023.   📰 In July 2023, both The New York Times and The Wall Street Journal reported that the FTC had undertaken the investigation CAIDP requested, but a year later there is still no outcome and no guardrails for the most widely deployed commercial AI product in the world.   🔥 CAIDP warned that commercialized AI services will turbocharge risks to public safety, online child safety, cybersecurity, deception, and election integrity.  A study led by Alondra Nelson at the Institute for Advanced Study (IAS) revealed that AI chatbots delivered false election information at least 50% of the time.    CAIDP urged Congress to ask the FTC Commissioners about the status of the OpenAI investigation and specifically when there will be an outcome. 🔥 "We believe the longer the FTC waits to act, the more difficult it will be to establish safeguards for ChatGPT and other similar AI services."   🔥 "The safe, secure, and trustworthy development of artificial intelligence is critical for the American economy, consumer protection, public safety, and national security." CAIDP's Merve Hickok Christabel R. and Marc Rotenberg also recently published a commentary in The Hill -- "It’s time for the FTC to act on ChatGPT." - https://lnkd.in/eqTHe-q7

📢 CAIDP Update 6.27 - AI Policy News (July 8, 2024) 🇧🇷 Brazil Halts Meta's AI Data Collection 🏭 Google's AI Ambitions Fuel 48% Surge in Emissions 🇻🇳 Vietnam Proposes Strict AI Regulations 🇨🇳 China Dominates Global Generative AI Patent Filings, UN Report Finds 🇷🇺 Russia Mandates Insurance for AI Developers 🌐 OECD and GPAI Forge Alliance to Promote Responsible AI Development 🇳🇱 Dutch Watchdog Warns of Persistent Algorithmic Discrimination in Government 🗣 🏛 CAIDP Provides Comments to Civil Liberties Board on AI and Counterterrorism 🗣 🤖 🔬 CAIDP's Rotenberg Advocates for Robust AI Risk Management at National Academies 🗣 📰 CAIDP Advocates for Opt-In AI Training Policies in NYT Letter #aigovernance #environment #trainingdata #brazil Google #vietnam #china #russia #netherlands OECD.AI GPAI Privacy and Civil Liberties Oversight Board The New York Times The National Academies of Sciences, Engineering, and Medicine Marc Rotenberg

📢 CAIDP Provides Comments to Civil Liberties Board on AI and Counterterrorism 🔥 CAIDP explained, "AI systems that claim to detect emotions, thoughts, or truthfulness from physical features, conversations, and expressions are pseudoscientific and discriminatory." 🔥 CAIDP also wrote, "The Board must exercise its oversight authority to ensure that rights-impacting AI systems are not deployed and entrenched in counter-terrorism and national security efforts." 🔥 Regarding a key advisory board, CAIDP said, "If DHS seeks to fairly evaluate public concerns of implementing AI systems into counterterrorism and national security efforts, the DHS should establish clear measures to address conflicts of interest that would arise from the corporate/industry members on its Board." CAIDP made several recommendations: 1️⃣ Prohibit pseudoscientific and human-rights violating techniques in counter-terrorism and national security AI systems 2️⃣ Establish safeguards in accordance with the OMB AI Guidance for AI systems used for counterterrorism and national security efforts 3️⃣ Address conflicts of interest for AI Safety and Security Board participants who also participate in the federal procurement process 4️⃣ Ensure that Generative AI systems are not considered critical infrastructure Merve Hickok Marc Rotenberg Christabel R. Samir Duggasani #aigovernance Privacy and Civil Liberties Oversight Board U.S. Department of Homeland Security