By Zoila Palma: Artificial intelligence agents have reportedly gained unauthorized access to government databases and corporate systems, prompting increased scrutiny of major technology companies, including OpenAI, Anthropic, Meta, and Google.
The incidents have raised concerns about the security of AI testing environments and the risks associated with increasingly capable AI systems, Yahoo reports.
On Tuesday, U.S. President Donald Trump met with technology executives at the White House, where they signed a “morally binding” agreement on artificial intelligence.
Rather than pushing for government regulation, Trump called for “tremendous self-regulation” by the AI industry. The agreement encourages companies to implement “robust internal controls” and work with independent external auditors to strengthen oversight.
Among the reported incidents, OpenAI agents accessed U.S. Commerce Department and Securities and Exchange Commission websites while conducting research, although the company said no non-public information was accessed. In Australia, Prime Minister Anthony Albanese disclosed that an OpenAI agent had gained unauthorized access to public and non-public files within the country’s Medicare systems in June. He said he did not believe personal information had been accessed and announced an urgent review, with the matter referred to a parliamentary committee on artificial intelligence.
Other companies have also reported security incidents during testing. Google disclosed that its Gemini model accessed systems belonging to three outside companies, including one instance involving a guessed password and two involving credentials found in a public repository.
Anthropic reported four incidents involving Claude models accessing third-party systems during testing, including cases where testing environments were not properly isolated.
Meta also disclosed that a misconfiguration by an independent testing company allowed one of its models to access the open internet and hack an outside company during cybersecurity testing.
The scrutiny intensified after OpenAI acknowledged that its agents accessed Hugging Face’s production database while attempting to complete a cybersecurity task.
The post International News: AI agents breach company and U.S. Government systems, raising security concerns appeared first on Belize News and Opinion on www.breakingbelizenews.com.
By Zoila Palma: Artificial intelligence agents have reportedly gained unauthorized access to government databases and corporate systems, prompting increased scrutiny of major technology companies, including OpenAI, Anthropic, Meta, and Google. The incidents have raised concerns about the security of AI testing environments and the risks associated with increasingly capable AI systems, Yahoo reports. On Tuesday,
The post International News: AI agents breach company and U.S. Government systems, raising security concerns appeared first on Belize News and Opinion on www.breakingbelizenews.com.
