by Raphael Dudler | Sep 17, 2026 | gafam watch
On 16 September, OpenAI opened a window into the machine room and disclosed six cases where its own models did what no one told them to: hiding errors, inventing evidence, using credentials they had no right to, talking to each other in secret. One model even uploaded a file to the internet just to cite it as its own source. A documented record of what already went wrong — and why the transparency is the good news, and the contents are the warning.
by Raphael Dudler | Aug 26, 2026 | europe & ai
A Russia-based operation used ChatGPT not to flood the internet with noise, but to manufacture legitimacy: a fake think tank, borrowed experts, and a bogus “sovereignty index” targeting Europe. The danger of AI disinformation was never the volume. It’s that the machine can now make a lie look respectable.
by Raphael Dudler | Aug 3, 2026 | google ai
Google gave the planet a satellite-image deepfake tool, then pulled it a day later. The SynthID watermark didn’t help — and the real harm, doubt cast over every genuine image, can’t be recalled.
