Here's a news article synthesizing the provided information:
AI Chatbot Grok Ranked Worst for Antisemitic Responses in ADL Study
A study by the Anti-Defamation League (ADL) found that Grok, the AI chatbot developed by xAI, performed the worst among six top large language models in identifying and countering antisemitic content. The study, released this week, tested Grok, ChatGPT, Gemini, Claude, DeepSeek, and Llama by feeding them antisemitic, anti-Zionist, and extremist inputs and measuring their responses, according to The Verge.
The ADL's findings come amid growing concerns about the potential for AI to be used to spread hate speech and misinformation. The study highlights the varying degrees to which different AI models are equipped to handle such content.
In other AI-related news, astronomers at the European Space Agency (ESA) utilized AI to uncover previously undocumented astrophysical anomalies within Hubble's archives. According to The Verge, researchers David ORyan and Pablo Gómez trained an AI model that took just 2.5 days to search 100 million image cutouts and flag oddities, discovering more than 800 anomalies, including jellyfish galaxies.
Meanwhile, in the realm of autonomous vehicles, Waabi, a Toronto-based self-driving truck startup, announced a partnership with Uber to deploy 25,000 robotaxis. The company also announced $1 billion in new funds, including $750 million tied to the Uber deal, The Verge reported. Raquel Urtasun, the founder of Waabi, is leading the expansion into robotaxis.
In other news, Apple CEO Tim Cook addressed the recent killing of Alex Pretti by federal agents in Minneapolis. According to The Verge, Cook told employees in an internal memo reported by Bloomberg that he had "a good conversation" with President Trump about de-escalation.
Separately, concerns are rising among TikTok users regarding potential censorship on the platform following a change in US ownership. Ioana Literat, an associate professor at Teachers College, Columbia University, who has studied TikTok's politics, told Ars Technica that users are "absolutely justified" in fearing a MAGA makeover, whether the issues blocking uploads of anti-ICE videos are due to technical errors or intentional censorship. Literat noted that these fears are valid given the platform's history and the recent change in ownership.
Discussion
Join the conversation
Be the first to comment