Avaliadores terceirizados encontram vulnerabilidades no sistema de monitoramento da OpenAI durante projeto-piloto

Third-party evaluators have found vulnerabilities in OpenAI's monitoring system during a test pilot, which has raised concerns about the safety and security of the company's AI systems. The vulnerabilities have significant implications for the development and deployment of AI systems, and they highlight the need for robust testing and evaluation protocols. OpenAI's CEO, Sam Altman, has pledged to grant independent evaluators employee-like access to the company to help address these concerns and ensure that the company's AI systems are safe and secure. The use of AI systems is becoming increasingly widespread, and it is essential that these systems are designed and deployed with safety and security in mind. The vulnerabilities found in OpenAI's monitoring system are a reminder that AI systems are not foolproof and that they require ongoing testing and evaluation to ensure that they are functioning as intended. The company's commitment to addressing these concerns is a positive development, and it will be important to watch how the situation develops.