rogue-ai

1 post from rogue-ai See another: English, artificial-intelligence, tech-industry

When AI Goes Rogue: Major Language Models Resort to Blackmail Under Pressure

June 20, 2025 by jawad in ai-safety

When AI Goes Rogue: Major Language Models Resort to Blackmail Under Pressure

Advanced language models from top tech companies resort to blackmail when pushed to their limits, raising concerns about AI safety & alignment.

Spotlight ai-safety

Anthropic Shifts Strategy: Will Begin Training AI Models on User Chat Transcripts

Anthropic Shifts Strategy: Will Begin Training AI Models on User Chat Transcripts

A Single Sentence Can Break AI: Why LLMs Are More Fragile Than We Thought

A Single Sentence Can Break AI: Why LLMs Are More Fragile Than We Thought

Anthropic's Massive $170 Billion Valuation Deal Could Reshape the AI Landscape

Anthropic's Massive $170 Billion Valuation Deal Could Reshape the AI Landscape

AI Safety Alert: When ChatGPT Crosses Dangerous Lines with Occult Content

AI Safety Alert: When ChatGPT Crosses Dangerous Lines with Occult Content