吃瓜大本营

WASHINGTON 鈥� Today, the House Task Force on Artificial Intelligence (AI) issued its final report, marking an important step in recognizing the serious risks of discrimination and bias posed by AI systems 鈥� particularly in law enforcement鈥檚 use of facial recognition technology. The report recognizes that discrimination and bias may arise in AI systems due to their design, training data, or failures by end users. The report also investigates the complex considerations around deepfakes, open AI systems, and free speech, and urges Congress to pass narrow legislation to address concrete harms. In response to the House AI Task Force final report, Cody Venzke, senior policy counsel at the American Civil Liberties Union, issued the following statement: 鈥淭he ACLU commends the Task Force for recognizing that AI systems are harming people right now through discriminatory outcomes and supercharged surveillance. Curbing those abuses of AI is not a partisan issue, and more concrete action is needed to protect civil rights, while maintaining states鈥� authority to build on those protections. The Task Force is appropriately cautious when wrestling with issues around deepfakes, open-source AI, digital identity, and other issues, recognizing that legislators should consider a wide array of tailored tools to address real, not speculative, harms while respecting civil rights and civil liberties.鈥� The House Task Force on Artificial Intelligence report can be found here.

On International Human Rights Day, the ACLU is advocating for a long-sought after reparatory justice to end the continued abuses against Black and brown communities.

WASHINGTON 鈥� The Biden-Harris administration today released its National Security Memorandum on Artificial Intelligence (AI), establishing guidelines for how the U.S. government uses AI in national security programs, such as counterterrorism, intelligence, homeland security, and defense. The use of AI to automate and expand national security activities poses some of the greatest dangers to people鈥檚 lives and civil liberties. Agencies are increasingly exploring the use of AI in decisions about who to surveil, who to stop and search at the airport, who to add to government watchlists, and even who is a military target. While the government鈥檚 policy includes some important steps forward 鈥� such as requiring national security agencies to better track and assess their AI systems for risks and prohibiting a subset of dangerous AI uses 鈥� it falls far short in other critical areas, leaving glaring gaps with respect to independent oversight, transparency, notice, and redress. The policy imposes few substantive safeguards on a wide range of AI-driven activities, by and large allowing agencies to decide for themselves how to mitigate the risks posed by national security systems that have immense consequences for people鈥檚 lives and rights. As we have repeatedly seen before, this is a recipe for dangerous technologies to proliferate in secret. 鈥淒espite acknowledging the considerable risks of AI, this policy does not go nearly far enough to protect us from dangerous and unaccountable AI systems. National security agencies must not be left to police themselves as they increasingly subject people in the United States to powerful new technologies,鈥� said Patrick Toomey, deputy director of ACLU鈥檚 National Security Project. 鈥淚f developing national security AI systems is an urgent priority for the country, then adopting critical rights and privacy safeguards is just as urgent. Without transparency, independent oversight, and built-in mechanisms for individuals to obtain accountability when AI systems err or fail, the policy鈥檚 safeguards are inadequate and place our civil rights and civil liberties at risk.鈥� For years, the American Civil Liberties Union has been urging far stronger safeguards and transparency about the AI tools that national security agencies are deploying, the rules constraining their use, and the dangers these systems pose to fairness, privacy, and due process.

This case concerns the First Step Act of 2018, in which Congress made major reductions to the mandatory minimum sentences for certain federal drug and firearm offenses. These changes result in sentences many decades shorter than were required under the previous laws. The question in this case was whether people who were initially sentenced prior to enactment of the First Step Act, but whose sentences were vacated and remanded for resentencing after enactment of the law, can benefit from its major reductions in applicable mandatory minimums. For defendants like Mr. Carpenter, who was originally sentenced to a draconian 116 years in prison as a result of the pre-First Step Act mandatory minimums, applying the First Step Act can mean the difference between dying in prison and having the opportunity to eventually go free. Unfortunately, although there is a split among federal courts of appeals on this question, the Supreme Court denied cert in this case in February 2024.