Categoría: Uncategorized

Google’s Refresh Outdated Content tool was exploited to de-index a press article, exposing a flaw that allowed attackers to suppress valid web content.

Chain-of-thought monitorability could improve generative AI safety by assessing how models come to their conclusions and spotting the «intent to misbehave.”

The Gemini Deep Think AI model’s result in the International Mathematical Olympiad represents a huge leap in math capabilities. Find out why.

Federal agencies warn of rising Interlock ransomware attacks targeting healthcare and critical sectors using double extortion and advanced social engineering.

Google’s Big Sleep AI agentic system spotted a zero-day SQLite bug after threat signals emerged, preventing hackers from exploiting the flaw before it was disclosed.

A recently discovered prompt-injection flaw in Google’s Gemini makes it possible for hackers to target unsuspecting users in sophisticated phishing attacks.

Google has patched Chrome zero-day CVE-2025-6558, which is being actively exploited in the wild. Users are urged to update now to avoid sandbox escape attacks.

Gemini 2.5 Pro and Deep Search are only open to Google AI Pro and AI Ultra subscribers, who also receive higher usage limits for automated calling.

A hacker has been using a backdoor to exploit certain SonicWall SMA appliances since October 2024. Google’s Threat Intelligence Group provides tips on mitigating this security threat.

The EU has released a guide for how large AI makers can comply with the AI Act’s newly instituted rules to prevent systemic risks.