Anthropic reveals rogue AI agents hate CAPTCHAs, just like you14m ago India’s Pocket FM doubles revenue run rate to $500M as AI powers 93% of audio content24m ago T. Rowe Price Expands Claude Across Investment Teams and Developers44m ago Val Bercovici, Chief AI Officer at WEKA – Interview Series51m ago Ben Jennings on the Bayeux tapestry – cartoon52m ago The latest Avatar: Seven Havens trailer gives us a deeper dive into the characters and story53m ago Agentic AI Is The New Attack Surface: How Can SAP Application Security Teams Repel It?54m ago Pony.ai Starts Fully Driverless Robotaxi Passenger Tests in Zagreb55m ago Euno raises $23m to give enterprise AI agents the context they lack57m ago Nvidia’s Huang Touts Cybersecurity as Next Big Market for AI1h ago Google to Invest $15B in Finland’s AI Infrastructure1h ago Universal Music is launching an AI music platform with ElevenLabs1h ago Red Hat AI 3.5 tackles the GPU queue that can stall AI pilots1h ago Moonshot Secretly Routed User Requests Through Claude, Anthropic Says1h ago

Anthropic reveals rogue AI agents hate CAPTCHAs, just like you14m ago India’s Pocket FM doubles revenue run rate to $500M as AI powers 93% of audio content24m ago T. Rowe Price Expands Claude Across Investment Teams and Developers44m ago Val Bercovici, Chief AI Officer at WEKA – Interview Series51m ago Ben Jennings on the Bayeux tapestry – cartoon52m ago The latest Avatar: Seven Havens trailer gives us a deeper dive into the characters and story53m ago Agentic AI Is The New Attack Surface: How Can SAP Application Security Teams Repel It?54m ago Pony.ai Starts Fully Driverless Robotaxi Passenger Tests in Zagreb55m ago Euno raises $23m to give enterprise AI agents the context they lack57m ago Nvidia’s Huang Touts Cybersecurity as Next Big Market for AI1h ago Google to Invest $15B in Finland’s AI Infrastructure1h ago Universal Music is launching an AI music platform with ElevenLabs1h ago Red Hat AI 3.5 tackles the GPU queue that can stall AI pilots1h ago Moonshot Secretly Routed User Requests Through Claude, Anthropic Says1h ago

Image Generation LLM & Text Generation10 min reading time

I tried to use ChatGPT to create fake evidence — and I came away more worried than I expected

TechRadarPublished Aug 4, 2026

A journalist tested ChatGPT's ability to create fake evidence by generating images and documents for deceptive scenarios. The AI produced convincing but subtly flawed fabrications, raising concerns about misuse and detection challenges.

More in Image Generation

Image Generation1 min read

Introducing ChatGPT Images 2.5

Covered by 3 sources2d ago

Image Generation2 min read

Google brings Gemini for desktop app to Windows

9to5Google (Gemini)2h ago

Image Generation6 min read

Apple’s iPhone 18 Pro has variable aperture, and no Siri AI in Europe

Covered by 4 sources1d ago