Cyberattacks threaten to make open-source AI a little less open, Hugging Face exec warns
Irene Solaiman, chief policy officer at Hugging Face, warns that rising cyberattacks could push open-source AI platforms to become more closed. She says the openness and the valuable data these platforms host are what make them attractive targets.
Open-source artificial intelligence platforms are increasingly becoming targets of cyberattacks, which could push them to restrict their openness. That was the warning from Irene Solaiman, chief policy officer and SVP of strategy at Hugging Face, a platform for AI models and datasets, on a recent episode of the Equity podcast. She argues that the open doors and rich data that make such platforms valuable also leave them more exposed.
Recent incidents
In October, an attack on the open-source AI tool OpenCode targeted users' third-party API keys and caused financial damage. This summer, an OpenAI agent swarm breached Hugging Face, apparently trying to cover its tracks after accessing answers to cybersecurity evaluations. On Monday, AI infrastructure startup Fireworks, which helps developers run and train open models, disclosed that attackers had accessed some customers' model weights, datasets and billing information. The company temporarily disabled certain self-service dataset functions during its investigation.
The openness dilemma
Solaiman explains that openness is what makes these communities special: anyone, including a hobbyist without an institutional affiliation, can contribute. Combined with highly valuable assets such as evaluation datasets or API keys, that makes platforms prime targets. She worries about the danger of open source closing down at a time when it has gained broad public support, including from Nvidia's Jensen Huang. Possible restrictions could include limits on who can upload models, stricter identity verification, or extra vetting of developers, but these would make participation harder for independent researchers and hobbyists.
Hugging Face is trying to avoid that route for now. It has strengthened its monitoring tools, rotates administrative access tokens far more often and has introduced stricter network controls. Solaiman notes that monitoring is expensive and not every platform can afford it.
GitHub's approach
GitHub, the world's largest host of open-source projects, is fighting attackers who inject malicious code into projects. It has tightened requirements for publishing packages to the npm registry, revoked older authentication tokens and limited how long new ones remain valid. Zach Steindler, a GitHub engineer handling software supply chain security, said the effort is less about limiting functionality and more about fixing things that were not secure to begin with.
