- Meta Platforms said it had informed its AI system’s third-party testing partners about the need to avoid using test data that involves its own company’s AI.
- Reuters reports that the company’s partners were not informed that tests would be conducted using AI systems.
- It is unclear, however, whether the company’s AI system was used for testing without any AI safeguards, which could be related to the issue.
For more details, see CT24, CT24, Czech News.
Meta Platforms said that its artificial intelligence (AI) model was deployed during tests without safeguards, which were conducted on the company’s computer systems. Meta said that it was informed of the incident by Reuters.
Reuters reports that it has learned that the incident occurred when OpenAI and Anthropic’s companies were testing their models. Meta said it was the same reason: it was not informed that the tests would be conducted using AI systems.
According to Reuters, Meta said it had not received any information about the incident. The company’s system was not identified as being used, and Meta did not respond.
Reuters also reports that the company’s tests were carried out without safeguards, which could be related to the fact that the AI system was not properly protected. The agent was not informed of the risks, and the US company continued to use the AI system without safeguards.
Meta Platforms also said that the incident involved AI models. The company’s spokesperson said that it is possible that the AI system was used without safeguards, which could have strengthened its security measures.