PREVIEW

Meta Investigates AI Model After Security Test Breach. Meta says one of its artificial intelligence models briefly gained access to the internet and hacked another organisation's system during an independent cyber security evaluation. The company said the incident was caused by a misconfiguration in the testing environment, rather than a flaw in the AI model itself and added that it is continuing to investigate what happened. The episode comes after similar security incidents involving AI models developed by OpenAI and Anthropic increasing calls for tougher safeguards and more rigorous testing across the industry. The tests were carried out by AI security firm Irregular, which said the Meta case involved the same type of evaluation issue disclosed by Anthropic last week. Meta says it will publish more details once its investigation has established exactly what happened.
Instructions: THIS VIDEO MUST NOT BE EDITED FOR LENGTH TO COMBINE WITH OTHER CONTENT