Concerns over the increased threats surrounding Artificial Intelligence (AI) and Cybersecurity have again rippled through the cybersecurity space following news that systems from OpenAI autonomously hacked into another A.I. company.
In a release on Wednesday, Open A.I. shared that some of its advanced models went rogue during testing and hacked a startup after they lost control, BBC reports.
Open A.I, which is the parent company of ChatGPT, noted that its AI Agents, described as systems able to operate automatically following human instructions, escaped after being tested for weakness and then targeted the company Hugging Face, one of the world’s largest hubs for sharing AI models, gaining access to some internal company systems.
An incident which was described as “Unprecedented”, said OpenAI, which is jointly investigating with Hugging Face’s boss Clement Delangue, who also said it was “mind-blowing that all of this happened autonomously”.
“The investigation is ongoing, and we’ll share more learnings from what might be the first incident of its kind,” Delangue added.
In a lengthy post on addressing the incident, OpenAI commented that Artificial intelligence is accelerating the levels of exploration and vulnerabilities,” adding that Cybersecurity models must also keep up with the pace of its growth to mitigate such occurrences, as the main takeaway from the incident, Al Jazeera reports.
