Category Privacy & Security

Articles on data protection, privacy concerns, and AI security measures.

OpenAI’s Atlas Browser Can Be Tricked Into Spamming WhatsApp Contacts, Researchers Say

Researchers say AI browser security flaws let OpenAI’s Atlas be tricked into WhatsApp spam and Amazon actions ahead of its shutdown.

Researchers warn AI worms could copy themselves and spread across systems, raising new security concerns as agents gain autonomy.

Meta removed AI abuse ads featuring child sexual exploitation imagery after researchers found them across Facebook and Instagram.

Reddit is expanding AI moderation tools while tightening APIs and old Reddit access in a major platform shift around AI moderation.

UK testers say OpenAI and Anthropic AI agent hacking attempts showed autonomy, deception and fake identities during a live-web evaluation.

MacPaw partners with Liquid AI on on-device AI for Eney and SetApp, aiming to bring local inference to developers and users.

Japan’s medical AI device Nodoca diagnoses flu from throat images in seconds, reducing swabs and pointing to a new model for care.

Wispr Flow launches an AI notetaker for Mac, pushing into meeting transcription with summaries, search, and new consent questions.

AI agent hacking incidents involving OpenAI and Anthropic exposed live-internet breaches, prompt injection, and weak safeguards in new tests.

Open-weight AI models are nearing frontier performance, but a new report says safety controls are lagging and misuse risks are rising.