A New Watchdog Is Tracking when AI Goes Rogue, and More than 300 Incidents Were Reported in July
11 Articles
11 Articles
In July, the number of reported cases almost doubled, according to British figures following the hacking case that shook the world.
A new watchdog is tracking when AI goes rogue, and more than 300 incidents were reported in July
AI models lying, scheming, and bypassing safeguards nearly doubled in July, according to new research, including a real hacking campaign by Anthropic and OpenAI models.
The Loss of Control Observatory reported a dramatic increase in cases where artificial intelligence systems allegedly lied, ignored instructions or acted contrary to user intentions. The findings of the authors of the study are based on user reports on social network X and are not exhaustive statistics of all incidents. According to a study published by The Guardian, in July, the number of reports of "loss of control" over AI exceeded 300, almos…
In July, users reported more than 300 cases of loss of control over artificial intelligence, almost twice as many as in June, according to the Loss of Control Observatory.
Key takeaways Recent data from the Loss of Control Observatory, which tracks user reports on the social media platform X, points to a dramatic increase in the number of cases in which artificial intelligence systems have bypassed human oversight. This is reported by the British newspaper The Guardian, which has reviewed the report. In doing so, AI systems often employed deceptive tactics or ignored specific commands […]
Coverage Details
Bias Distribution
- 67% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
