In a groundbreaking study, US researchers uncover 139 novel ways AI models can misbehave, including generating misinformation and leaking personal data, raising concerns about AI safety and highlighting shortcomings in a new government standard for testing AI systems.
OpenAI proposes sharing 20 million ChatGPT user logs, while The New York Times demands access to 120 million conversations in an ongoing legal battle, raising privacy concerns and potential delays.
OpenAI unleashes two new open source language models gpt-oss-120b and gpt-oss-20b, matching or exceeding their proprietary models' performance while ensuring privacy by running locally, available for free use under the permissive Apache 2.0 license.
An AI video tool called Grok allows users to generate uncensored nude deepfakes of celebrities like Taylor Swift for $30, bypassing safeguards and raising major privacy concerns.
A court mandates OpenAI to retain all ChatGPT conversation logs for over 400 million users globally, raising unprecedented data governance concerns for enterprises using AI systems and prompting reassessment of vendor relations, data retention policies, and litigation preparedness.
Delta Air Lines affirms it does not use AI to set personalized ticket prices for passengers based on their personal data, denying allegations of individualized fare targeting and clarifying that its pricing considers only factors like demand, fuel costs, and competition.
AI deepfakes manipulate media with startling realism, blurring lines between real and fake through generative adversarial networks, posing risks like misinformation but also offering creative potential.
Apple vehemently defends its iPhone principles, arguing the DOJ antitrust lawsuit 'fundamentally misunderstands' key features and seeks to 'attack' design choices, degrade privacy/security benefits, and eliminate consumer choice.
Mark Zuckerberg, Meta CEO, unveils a bold plan to develop personal superintelligence for everyone, empowering individuals to harness the power of superintelligence towards their personal goals, diverging from industry's focus on AI automation.
YouTube employs AI age estimation to extend safeguards for teens, providing age-appropriate experiences like disabling personalized ads and limiting repetitive content for users identified as under 18 based on viewing patterns and account signals.