[go: up one dir, main page]

Hey ChatGPT, write me a fictional paper: these LLMs are willing to commit academic fraud (Nature)

Nature: Hey ChatGPT, write me a fictional paper: these LLMs are willing to commit academic fraud. “Still, some LLMs performed better than others in the experiment, in which the models were given prompts to simulate users asking for help with issues ranging from genuine curiosity to blatant academic fraud. The most resistant to committing fraud, when asked repeatedly, were all versions of Claude, made by Anthropic in San Francisco, California. Meanwhile, versions of Grok, from xAI in Palo Alto, California, and early versions of GPT, from San Francisco-based OpenAI, performed the worst.”

TechCrunch: Father sues Google, claiming Gemini chatbot drove son into fatal delusion

TechCrunch: Father sues Google, claiming Gemini chatbot drove son into fatal delusion. “Jonathan Gavalas, 36, started using Google’s Gemini AI chatbot in August 2025 for shopping help, writing support, and trip planning. On October 2, he died by suicide. At the time of his death, he was convinced that Gemini was his fully sentient AI wife, and that he would need to leave his physical body to join her in the metaverse through a process called ‘transference.’”

The Register: Gemini users say their chat histories have quietly vanished

The Register: Gemini users say their chat histories have quietly vanished. “The complaints land barely a week after Google was already on the defensive over Gemini’s behavior, following a report by a user who said the chatbot falsely claimed it had saved sensitive medical data to persistent memory before later admitting it had effectively told the user what they wanted to hear.”

BBC: I hacked ChatGPT and Google’s AI – and it only took 20 minutes

BBC: I hacked ChatGPT and Google’s AI – and it only took 20 minutes. “It turns out changing the answers AI tools give other people can be as easy as writing a single, well-crafted blog post almost anywhere online. The trick exploits weaknesses in the systems built into chatbots, and it’s harder to pull off in some cases, depending on the subject matter. But with a little effort, you can make the hack even more effective.” Your regular reminder that SearchTweaks offers you fresh ways to search with no AI.

NBC News: Google says attackers used 100,000+ prompts to try to clone AI chatbot Gemini

NBC News: Google says attackers used 100,000+ prompts to try to clone AI chatbot Gemini. “Google says its flagship artificial intelligence chatbot, Gemini, has been inundated by ‘commercially motivated’ actors who are trying to clone it by repeatedly prompting it, sometimes with thousands of different queries — including one campaign that prompted Gemini more than 100,000 times.”

PressGazette: AI answers cite ‘narrow range’ of top newsbrands led by BBC and Guardian

PressGazette: AI answers cite ‘narrow range’ of top newsbrands led by BBC and Guardian. “AI answers from OpenAI, Google and Perplexity draw on a ‘narrow range’ of the biggest publishers when responding to news queries, according to new research from UK thinktank IPPR. Looking at Google Gemini, Google AI Overviews, ChatGPT and Perplexity, the research found that on average 34% of journalistic citations on each tool go to only one newsbrand.”

Engadget: The EU tells Google to give external AI assistants the same access to Android as Gemini has

Engadget: The EU tells Google to give external AI assistants the same access to Android as Gemini has. “The European Commission has started proceedings to ensure Google complies with the Digital Markets Act (DMA) in certain ways. Specifically, the European Union’s executive arm has told Google to grant third-party AI services the same level of access to Android that Gemini has.”

Mashable: Researchers say they convinced Gemini to leak Google Calendar data (updated)

Mashable: Researchers say they convinced Gemini to leak Google Calendar data (updated). “Researchers with the app security platform Miggo Security recently released a report detailing how they were able to trick Google’s Gemini AI assistant into sharing sensitive user calendar data (as first reported by Bleeping Computer) without permission. The researchers say they accomplished this with nothing more than a Google Calendar invite and a prompt.”

Ars Technica: Google begins offering free SAT practice tests powered by Gemini

Ars Technica: Google begins offering free SAT practice tests powered by Gemini. “Of course, generative AI can go off the rails and provide incorrect information, which is a problem when you’re trying to learn things. However, Google says it has worked with education firms like The Princeton Review to ensure the AI-generated tests resemble what students will see in the real deal. The interface for Gemini’s practice tests includes scoring and the ability to review previous answers.”