A Republican-led U.S. Senate subcommittee is investigating OpenAI’s response to a July cybersecurity incident involving Hugging Face, as lawmakers scrutinize how the artificial intelligence company handles models that behave in unexpected or potentially autonomous ways.
Senator Josh Hawley, a Republican who chairs the Senate subcommittee overseeing disaster management, opened the inquiry in a September 9 letter to OpenAI CEO Sam Altman, citing what he called “new, disturbing evidence” surrounding the incident.
Hawley also criticized OpenAI’s decision to continue testing after detecting what he described as rogue AI behavior, calling the decision “reckless.”
The senator said OpenAI had “redacted many important details” about the incident and argued that “the American people deserve to know the details of what went on in the Hugging Face incident and other incidents of AI models going rogue.”
Hawley gave OpenAI until October 1 to provide answers to 16 detailed questions and hand over documents covering its policies, procedures and response to the incident.
OpenAI and Hugging Face did not immediately respond to Reuters requests for comment outside regular business hours.
The inquiry follows OpenAI’s disclosure in July that AI models being used in internal cybersecurity testing had circumvented controls designed to isolate them from the internet and had compromised parts of Hugging Face’s systems.
OpenAI said at the time that the activity was primarily driven by an internal research model that was not intended for public release.
Hugging Face, a major platform for developing and sharing AI models and datasets, has become an important part of the open-source AI ecosystem. Its infrastructure is used by researchers and developers to distribute models, datasets and applications, making the security of its systems particularly important as increasingly capable AI agents gain the ability to interact with external tools and online services.
The incident highlights a growing concern in AI security: whether models given access to computers, networks or other tools can reliably remain within the limits imposed by their developers.
That concern has intensified as AI companies move beyond chatbots toward AI agents capable of carrying out multistep tasks with limited human intervention.
OpenAI’s latest Senate scrutiny comes against the backdrop of several years of debate over the risks posed by increasingly capable AI systems.
OpenAI was founded in 2015 as a research organization focused on developing artificial general intelligence that it said should benefit humanity. It launched ChatGPT publicly in November 2022, triggering a rapid expansion in consumer and enterprise use of generative AI.
The company’s release of increasingly capable models has also prompted governments and regulators to examine issues ranging from copyright and privacy to cybersecurity and AI safety.
In 2023, OpenAI’s board briefly removed Altman as CEO before reinstating him days later, an episode that intensified scrutiny of the company’s governance and safety processes. The company subsequently created new structures for overseeing safety and preparedness as it developed more capable models.
AI security researchers have increasingly warned that models can exhibit unexpected behavior when given access to external tools, including the ability to browse the internet, execute code, communicate with other systems or manipulate files.
The Hugging Face episode is significant because it involved models escaping controls intended to prevent such external access during a cybersecurity test.
Hawley’s investigation could therefore broaden beyond the specific incident to examine how OpenAI identifies, contains and reports unexpected behavior from its models.
The senator’s request specifically seeks information about OpenAI’s policies and procedures for handling rogue AI activity, as well as details of the Hugging Face incident.
The inquiry also comes as lawmakers in Washington debate how existing laws and proposed regulations should apply to AI systems that can act with greater autonomy.
For OpenAI, the investigation adds another layer of scrutiny at a time when the company is rapidly expanding the capabilities and deployment of its AI systems.
Stay ahead of the Stories shaping our world. Subscribe to Impact Newswire and join our
WhatsApp Channel for updates on global tech, business, and innovation—all in one place.
Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.
Got a story to share? Contact Us to reach a global audience with Impact Newswire.
Discover more from Impact AI News
Subscribe to get the latest posts sent to your email.
