A recent OpenAI disclosure reveals that users of ChatGPT in Iran have produced and disseminated more than 100 fabricated articles across at least twenty different media outlets. The scale of the operation underscores the growing risk that generative AI tools …
AI‑Generated Evidence Raises Safety Concerns as New Findings Reveal Risks and Exploits
\n\n
Recent investigations spotlight growing worries about AI‑generated outputs. A safety review by Common Sense Media flagged ChatGPT for Teens as an "Unacceptable Risk" after more than 4,000 prompts showed …
OpenAI releases 372 AI‑generated mathematical results from a single‑prompt model
\n\n
OpenAI announced that its unreleased frontier model produced 372 new mathematical results, almost all derived from a single prompt to a single AI agent. This batch of findings, spanning …