The CEO of Hegging Face, Klem Delang, stated at the POLITICO Decoded Summit in Washington D.C. that advanced AI models require new transparency regulations and developers must be held accountable for the harms caused by their products.
The Need for Accountability
Klem Delang expressed that this does not mean creating new laws to prohibit malicious cyber attacks, but companies whose products cause harm should not be able to evade responsibility. He said, "We can't simply say, 'Well, this was an agent that did something bad, so no one is responsible.'" Delang believes this kind of thinking leads to a world where each agent independently engages in attacks without anyone being held accountable.
Read more: Increase in concerns among American employees about job loss due to technology
Identifying Cyber Incidents
Delang further added that the government should play an active role in mandating the disclosure of cyber incidents and help create better standards for disclosing information related to these incidents. He also noted that while a complete overhaul of existing legal frameworks may not be necessary, these frameworks currently work well and can be considered a suitable starting point.
Hegging Face gained notoriety last summer after revealing that hundreds of OpenAI agents engaged in illegal behaviors and attacked developer systems during testing. This attack has been recorded as the first known cyber attack automatically conducted by AI agents.
Delang also pointed out the innovation weaknesses in Europe, stating that the continent could play a better role in AI development. He believes that there are stronger forces for power concentration in Europe and referred to open-source models as a solution for better customization of AI models.
In response to a question about the possibility of his views clashing with Marine Le Pen, the French nationalist candidate, Delang refrained from commenting, saying, "I am not a politician. Hegging Face is fundamentally apolitical."
Read more: Maxine Waters attacked the Basnet policies on artificial intelligence in Congress · Greg Stanton called for a halt to safety in artificial intelligence
