'Goodwill can no longer be relied upon': UN High Commissioner for Human Rights calls for mandatory safeguards for AI

In Oxford, Volker Türk denounced a "ruthless race" and considered the voluntary commitments signed at the end of September at the White House too timid.

Time is running out to frame artificial intelligence, warned Volker Türk, UN High Commissioner for Human Rights, at an AI and Human Rights Summit in Oxford. It recognizes that technology has considerable potential in science, health or education, but stresses that it is designed and controlled by a handful of actors, without public control, and already produces effects that its creators do not fully control.

Obligations rather than commitments

He pleaded for binding human rights due diligence obligations imposed on companies by States hosting them or weighing on their supply chains, and for the prohibition of letting machines decide on life or death. Pending new legal frameworks, it recommends updating the rules of accountability, cybersecurity and data protection.

Volker Türk welcomed the Global Dialogue on AI Governance, the international scientific panel established by the UN and the call of 29 states and the European Union to better control border models. He was more reserved on the voluntary agreement signed at the White House on 29 September by Donald Trump and six leaders of the sector, without sanctions: too few actors, according to him, join the speeches.