Sam Altman, CEO of OpenAI, has reignited the discussion on legal responsibilities in the field by stressing the need for accountability frameworks for AI companies. In an interview, he stated: "If something goes wrong in our models during training, we must take responsibility for it."
Reviving the Accountability Discussion
Altman's statements have prompted some influential members of the European Parliament to call for the revival of legislation on AI liability. This comes after the European Union had shelved plans last year to establish liability laws and compensation for damages caused by AI.
Alexis Foos, a German conservative representative who previously worked on this proposal in the European Parliament, reacted to Altman's comments, saying: "This is a late moment of honesty. For years, the tech industry has claimed that accountability can stifle innovation. Now even the CEO of OpenAI admits: Without accountability, there is no trust."
AI Liability Laws
Altman believes that there needs to be a law for companies' accountability regarding AI models. He pointed out that the EU's proposed framework in 2022 could have addressed this need by clarifying responsibilities before development, during training, and testing of models.
As concerns grow over the rapid advancement of AI and its potential harms, the issue of accountability has become one that cannot be overlooked. Some lawmakers in the European Union particularly feel that there is a strong need for new laws regarding accountability in this area.
Brando Benfi, a representative of the Social Democratic Party, also stated that abandoning the accountability proposal was a serious mistake and added: "Advanced models and high-risk systems require strict accountability." He and several other lawmakers are calling for the introduction of a law for AI accountability to address existing legal gaps in this area.
While the European Union is exploring new solutions to tackle the risks associated with AI, other countries are also closely monitoring these developments.
