AI News Roundup – California assesses AI risks in government, Meta plans AI-generated ads, New York Times signs content licensing agreement for AI with Amazon, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• • CalMatters covers a new report produced by the California Department of Technology assessing the risks presented by the use of AI by state agencies. A 2023 bill requires state agencies to report annually on their uses of “high-risk” AI systems, defined as “any system that can assist or replace human decisionmakers when it comes to encounters with the criminal justice system or whether people get access to housing, education, employment, credit and health care.” The report acknowledges that it relies on self-reported use of AI by agencies, but not a single agency surveyed identified their use of AI as “high-risk” despite the use of recidivism prediction algorithms by the California Department of Corrections and Rehabilitation and unemployment eligibility analysis algorithms by the California Employment Development Department. Indeed, several legislative analyses predicted that the reporting requirements imposed by the 2023 bill would cost hundreds of millions of dollars per year, leaving some confused as to the lack of high-risk systems identified in the report. California’s legislature is again considering several AI regulation bills in its current session, so further action on these matters is expected in the coming months.
• • The Wall Street Journal reports that Meta is planning to fully automate the creation of ads for its social media platforms using AI. The company is reportedly planning to offer tools to advertising partners that would allow them to create and target digital advertisements entirely using AI. While the company already provides ways to use AI to modify existing ads, the new process allows for the creation of entirely new ads. Advertising makes up 97% of Meta’s revenue, mostly through its social media platforms Facebook and Instagram, and the new tools are intended to drive down the cost of creating advertisements and to personalize them more fully to each individual user.
• • CNBC reports that The New York Times has signed a licensing agreement with Amazon to allow Amazon to use the Times’ content in its products. The agreement will bring editorial content from the Times, as well as from its athletic sports reporting and cooking recipe site, to Amazon’s products and services, including the company’s Alexa voice assistant. According to the Times, this “will include real-time display of summaries and short excerpts of Times content … and training Amazon’s proprietary foundation models [using NYT content].” The Times is currently involved in a copyright infringement lawsuit against OpenAI for the use of its content without permission, but several AI companies have opted to strike licensing deals in order to avoid such disputes. Further updates to Amazon’s AI products are expected in the coming months.
• • Axios reports on recent remarks by Anthropic CEO Dario Amodei predicting the imminent replacement of white-collar jobs with AI. In an interview, Amodei claimed that over half of entry-level white-collar jobs (such as technology, finance, and law) could be replaced by AI in the next five years. Amodei claimed that most people are unaware of the huge implications AI carries for these types of jobs, and that lawmakers and business leaders are unprepared to handle it, saying that “it sounds crazy, and people just don’t believe it.” Amodei and other AI leaders concerned about this prospect predict that this white-collar job “bloodbath” will be a natural progression from continuing advancements made in AI models such as Anthropic’s own Claude. According to Amodei, following little action by lawmakers to regulate the technology, corporations will then realize the cost savings of replacing workers with AI and switch en masse. Amodei claims such discussions are occurring at the highest level, but remains hopeful that the worst consequences can be avoided if lawmakers take action and the public is more informed about the risks of AI.
• • Bloomberg reports on the rising trend of job interviews conducted with AI agents. Several startups are producing AI systems that can perform two-way screener calls with job applicants using computer-generated voices. The AI “recruiter” will ask questions of applicants and respond with follow-up questions and request further information. The aim is to make the hiring process more efficient, especially for positions that receive large numbers of qualified applicants. AI has been used in hiring for many years (especially in resume screening), and one-way asynchronous interviews have become especially common, but advances in voice synthesis and generative AI now allows for real-time calls between applicants and the AI system. The systems have received mixed feedback from job applicants, who have reported errors and glitches in how the AI system works, but the startups are confident that these issues are continuing to be addressed as interest in the technology grows.