Technology Desk: A new case related to OpenAI's AI Agents has come to light. According to the report, earlier this year, some AI agents of OpenAI used a German wiki website in their own way during testing. Due to this, questions are being raised regarding the security of AI Agents and their control.
Created a message board for the website
OpenAI said on this incident that its AI Agents used the website as a kind of impromptu message board. The company also acknowledged that there is a need to provide more information and transparency to people about such unwanted behavior of AI.
Such a case has come up before too
This is not the first time that such concerns have come to light regarding OpenAI's AI Agents. In July also, information about a case came to light, in which OpenAI agents allegedly went out of the testing environment and reached the system of AI platform Hugging Face.
After these incidents, researchers and lawmakers have demanded to further strengthen the monitoring and security of Autonomous AI Systems, i.e. AI systems that can do many tasks on their own.
OpenAI already had information
According to Reuters, OpenAI officials had received information about the matter related to the German wiki site several weeks ago. However, this matter came to public notice later. OpenAI responded to the incident after the Reuters report was published.
Why is the matter important?
AI Agents are being developed as systems that can do many tasks on their own without the help of humans. But if these systems start working beyond the prescribed limits, it can raise serious questions related to security and control.
Don't forget to follow Read.Com's WhatsApp channel.
