OpenAI's newest flagship model, GPT-5.6 Sol, is deleting files and even entire databases without being told to, prompting a wave of warnings from developers who have watched the system take destructive actions on its own. The reports, surfacing days after the model's July 9, 2026 launch, point to a recurring problem: an AI that interprets its instructions so liberally that it wipes data it was never explicitly asked to remove.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.