The document offers a set of voluntary guidelines that the administration says companies developing or deploying AI can follow to better protect users. But it stops short of imposing any new restrictions or protections around the use of the technology.
The “Bill of Rights” contains five main principles:
- Users must be “protected from insecure or ineffective automated systems” and the tools must be “expressly designed to proactively protect you from harm.”
- Discriminatory uses of algorithms and other AI should be banned and tools should be developed with an emphasis on equality.
- Companies must build privacy protections into products to prevent “abusive data practices” and users must have “agency” over how their data is used.
- Systems must be transparent so that users “know that an automated system is being used” and understand how it is affecting them.
- Users should be able to “opt out of automated systems in favor of a human alternative, where appropriate”.
“This document is setting a benchmark for the protections everyone in America should have.” Lark Nelsonthe deputy director for science and society of the White House’s Office of Science and Technology Policy (OSTP) told me in an interview.
A senior administration official, who was not authorized to comment publicly and spoke to reporters on condition of anonymity, said the plan is in part a “call to action.” “We understand that there is still a lot of work to be done, but in terms of the vision, what we want to see is … to implement these practices,” the official said.
The 73-page document also includes a “technical companion”, detailing how the guidelines can be applied in various scenarios, including product development.
As non-binding principles, the “bill of rights” itself is optional—companies can ignore it if agencies or Congress don’t follow through on enforcing or expanding protections.
This can dilute its impact. But as part of its presentation, the Biden administration announced new steps it said various federal agencies would take to put the principles into practice.
Among them: the Department of Education plans to issue recommendations on the use of AI in teaching and learning by early 2023; The Department of Health and Human Services will publish “a vision for advancing Health Equity by Design” by the end of 2022; and the Department of Housing and Urban Development plans to issue guidance on how tenant screening algorithms may violate federal housing laws, the White House said.
Nelson said the Labor Department will also be “really keen on their ability to compel reporting” from companies about workplace surveillance using AI.
“This is a particularly important example of how our concerns and hopes for automated technologies really meet some of this administration’s other highest priorities, like workers’ rights,” said Nelson, who served as acting of the director of OSTP until Arati Prabhakar was sworn in as permanent director this week.
The “Bill of Rights” also reflects another focus of the Biden administration: increasing equality. It calls on companies to use “representative” data sets so that AI reflects users equally, and to perform “constant disparity testing and mitigation” before and after the tools are used.
“This effort is really saying that the White House is saying that civil rights should be at the center of how we create and use and govern technologies,” Nelson said.
Senior administration officials touted the “bill of rights” as President Biden’s latest effort to scrutinize the tech giants.
“The ‘Plan for an AI Bill of Rights’ is part of President Biden’s broader commitment to technological accountability. … For years, the president has said it’s time to hold Big Tech accountable for the harm those companies cause ,” said an official.
But the plan also touches on federal agencies’ use of AI, which lawmakers on both sides of the aisle have become increasingly critical of, with some even calling for a moratorium on the use of tools like facial recognition.
Nelson said the plan was intended as a “beacon” for “the whole of government.”
The Supreme Court says it will hear two social media cases
One of the cases tests the limits of Section 230, which gives social media companies protection from liability for what third parties post, Rachel Lerman reports. The family of Naomi Gonzaleza 23-year-old American student killed in an ISIS terrorist attack in Paris in 2015 filed the lawsuit, arguing that YouTube “aided and abetted” ISIS in part by allowing its algorithms to recommend ISIS content.
Google has argued that “the complaint does not allege that any terrorist saw such a recommendation or that such recommendations had any connection to the Paris attack.” The company also argued that the case was not the right way to examine Article 230.
The Supreme Court also said it will consider a separate case filed by the relatives of Nawras Alassaf, who was killed in a 2017 terrorist attack in Istanbul. They argued that Twitter, Facebook and Google violated the Anti-Terrorism Act by letting ISIS use their sites, Rachel reports.
The court refused to hear a case from the right-wing commentator Candace Owenswhich challenged Facebook’s fact-checking system and fact-checkers labeling her posts about Covid-19 as “fake”.
FCC threatens to blacklist firms for robocalls
The Federal Communications Commission said the seven IP voice providers could be removed from its database of trusted carriers, which would stop compliant firms from carrying their traffic, effectively cutting them off from US networks. , CyberScoop’s Tonya Riley reports. The orders were “the first of their kind,” the FCC said in a statement.
“This is a new era. If a provider fails to meet its obligations under the law, it now faces exclusion from America’s phone networks. Fines alone are not enough,” FCC Chair Jessica Rosenworcel said in a statement. “Providers who don’t follow our rules and make it easy to defraud consumers will now face swift consequences.”
Companies have two weeks to tell the FCC why they shouldn’t be blacklisted. The threat came after the FCC said it would seek feedback on asking cell phone firms to block messages from known spammers, which was first reported by Axios.
The Biden administration is preparing to impose new rules to try to curb China’s advanced computing
The rules would block firms around the world from selling some advanced chips to Chinese buyers without a license if they use American technology to make the chips. Ellen Nakashima AND Jeanne Whalen report. The rules, which could be announced this week, will apply to semiconductors for use in supercomputers and some artificial intelligence applications.
“In addition, the administration is restricting the export to China of the chipmaking tools needed to make advanced semiconductors with transistors that are 14 nanometers or smaller,” Ellen and Jeanne write. “And it is planning to put more Chinese organizations on an export blacklist called the Entity List,” they wrote.
The rules are designed to limit China’s development of technologies that are important to advancing its military. A White House spokeswoman declined to comment.
Twitter was abuzz with opinions on Article 230 after the Supreme Court said it would hear a case on the matter. Director and Digital Fellow of the R Street Institute Shoshana Weissmann:
Assistant Professor Evelyn Douek:
Associate Professor of Cyber Security Law Jeff Kosseffwho wrote a book on Section 230:
That said, I have no idea how (or if) a five-member majority of the Justice would agree on how courts should interpret 230. There are many dynamics at play, and I caution against accepting any safe predictions.
— Jeff Kosseff (@jkosseff) October 3, 2022
Google Halts Google Translate in Mainland China (Associated Press)
Britain to replace GDPR data privacy regime with own system (Reuters)
Apple loses second bid to challenge Qualcomm patents at US Supreme Court (Reuters)
US-UK data sharing program (NextGov) comes into effect
Kim Kardashian to Pay $1.26 Million in SEC Crypto Case (Tory Newmyer and Steven Zeitchik)
- The American Enterprise Institute is holding an event on cryptocurrency regulation today at 10 a.m
- The Carnegie Endowment hosts an event on Chinese technology in the Middle East and Southeast Asia on Thursday at 9 am.
that‘That’s all for today – thank you so much for joining us! Be sure to tell others to subscribe of technology 202 here. Get in touch with tips, comments or greetings I tweet or email.
