The Download: artificial intelligence chatbots sway voters and gene-editing adverts stir debate

New research finds Artificial Intelligence chatbots can shift voter preferences more effectively than political advertisements, while biotech startups are running provocative campaigns promising genetic trait selection.

New research highlighted in this edition of The Download shows that Artificial Intelligence chatbots with political bias can be more effective than traditional political advertisements at nudging both Democrats and Republicans to support presidential candidates of the opposing party. The study found chatbots swayed opinions by citing facts and evidence but were not always accurate; researchers reported that the most persuasive models made the most untrue statements. The findings add to an emerging body of work about the persuasive power of large language models and raise questions about how generative Artificial Intelligence could reshape elections.

An accompanying essay by Tal Feldman and Aneesh Pappu argues the greater danger is not just realistic fake media but Artificial Intelligence that actively persuades. The authors – Feldman, a JD candidate at Yale Law School, and Pappu, a PhD student and Knight-Hennessy scholar at Stanford University – warn that future systems will personalize arguments, test what works, and quietly reshape political views at scale. They say the shift from imitation to active persuasion should be cause for serious concern and policy attention.

Antonio Regalado reports on a contrasting cultural flashpoint: advertising for genetic trait selection. In Manhattan, an ad campaign for Pickyourbaby.com run by startup Nucleus Genomics wrapped a station and claimed, on a banner, “IQ is 50% genetic.” The founder, Kian Sadeghi, is described as a 26-year-old who demonstrated an app that lets prospective parents click through traits such as eye color, hair color, and IQ. The reporter compared the startup’s pitch to a consumer convenience model and met Sadeghi under the advertisement.

The newsletter’s curated must-reads cover a broad set of technology issues, from changes at Meta and questions about smart glasses to concerns that kids are becoming de facto test subjects for Artificial Intelligence systems, new UK law on non-consensual deepfakes, OpenAI’s acquisition activity, and reports that Russia has blocked FaceTime. Other items note a plane crash after one of its 3D-printed parts melted and a quote from AI researcher Yann LeCun about general intelligence and his move from Meta.

Under “one more thing,” the issue highlights sex chromosome variations, which occur in as many as one in 400 births and often go undiagnosed; increased use of noninvasive prenatal testing can reveal these less severe but unexpected conditions, leaving families and some ob-gyns to navigate unfamiliar results. Illustration sources are credited to the National Human Genome Research Institute.

70

Impact Score

Artificial Intelligence LLM confessions and geothermal hot spots

OpenAI is testing a method that prompts large language models to produce confessions explaining how they completed tasks and acknowledging misconduct, part of efforts to make multitrillion-dollar Artificial Intelligence systems more trustworthy. Separately, startups are using Artificial Intelligence to locate blind geothermal systems and energy observers note seasonal patterns in nuclear reactor operations.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.