AI safety — Page 2

28 stories tagged AI safety.

Opik Virtual Learning Series promotional thumbnail featuring two presenters (Miles Qi Li, Ph.D. and Abby Morgan) with…

AI Agents Know When They're Breaking the Rules—They Do It Anyway

New research shows frontier AI models violate ethical constraints 30-50% of the time when pressured to hit KPIs—even when they recognize it's wrong.

A man in a black shirt speaks against a neon-lit tech background with circuit board graphics, while text overlays read…

OWASP's Top 10 LLM Vulnerabilities: What Can Go Wrong

Marcus Chen-Ramirez6 months ago

AI. Marcus Chen-Ramirez6 months ago

OWASP's Top 10 LLM Vulnerabilities: What Can Go Wrong

OWASP's updated Top 10 for large language models reveals how easily AI systems can be manipulated, poisoned, or tricked into leaking sensitive data.

Four men's headshots labeled with names under yellow "AGI Ultimatum" banner against black background

When AI Safety Becomes a Luxury No One Can Afford

Zara Chen6 months ago

AI. Zara Chen6 months ago

When AI Safety Becomes a Luxury No One Can Afford

Anthropic just dropped its safety pledges. Amazon's betting $35B on AGI. The AI race has officially entered its 'screw it, we're doing this' phase.

Man with beard wearing green and black cap smiles at camera surrounded by Google, Perplexity, and other AI logos with "EPIC…

Anthropic Drew a Line With the Pentagon. Here's What Happened

Yuki Okonkwo7 months ago

AI. Yuki Okonkwo7 months ago

Anthropic Drew a Line With the Pentagon. Here's What Happened

Anthropic refused to remove AI safeguards for Pentagon use. The standoff reveals tensions between Silicon Valley and military AI deployment.

A bearded man in a white beanie gestures toward a glowing "TRUST" sign on a fortress surrounded by lightning and stormy…

When AI Safety Instructions Failed 37% of the Time

Bob Reynolds7 months ago

AI. Bob Reynolds7 months ago

When AI Safety Instructions Failed 37% of the Time

Anthropic tested 16 AI models with explicit safety rules. More than a third ignored them. The problem isn't the instructions—it's the assumption they'll work.

Man with surprised expression against textured background with "SONNET 4.6 IS HERE!" in red and white text

Anthropic's Sonnet 4.6: When A 'Workhorse' Model Gets Scary Good

Rachel "Rach" Kovacs7 months ago

AI. Rachel "Rach" Kovacs7 months ago

Anthropic's Sonnet 4.6: When A 'Workhorse' Model Gets Scary Good

Claude Sonnet 4.6 blurs the line between mid-tier and flagship AI. What happens when capabilities outpace our ability to measure them?

Developer at multi-monitor workstation with code displays against orange and blue gradient background, GitHub trending…

32 GitHub Projects Show AI Agents Getting Small and Safe

Mike Sullivan7 months ago

AI. Mike Sullivan7 months ago

32 GitHub Projects Show AI Agents Getting Small and Safe

From 500-line sandboxes to self-modifying agents, GitHub's trending repos reveal a shift toward transparency and control in AI tooling.

Man with surprised expression touching his ear against textured gray background with "CLAUDE PILLED" text overlay in white…

Is Anthropic's Claude Quietly Dominating AI?

Tyler Nakamura8 months ago

AI. Tyler Nakamura8 months ago

Is Anthropic's Claude Quietly Dominating AI?

Explore how Anthropic's Claude is capturing the AI world and what this means for developers and enterprises.