Universities must lead efforts to ensure that the evolution of artificial intelligence tools does not discriminate against developing nations, a conference has heard.

Speaking on a panel at ̽����Ƶ’s Emerging Economies Summit, Sarah Anyang Agbor, commissioner for human resources, science and technology at the African Union, said that the need for global diversity had been overlooked in the development of machine learning technology.

“It is defined by the perspective of the West…the language of the machine is not the language of Africa,” she told the event at Qatar University. “Who is designing the machines that will dictate the future of all of us and do they come from a position of bias?”

Professor Anyang Agbor, who was previously a deputy vice-chancellor of the University of Bamenda, in Cameroon, said that universities must play a key role in correcting the omission of perspectives from the developing world.

̽����Ƶ

“It is the job of the university to make [AI] more inclusive and participatory. It can be done because we have the intellect to do it [and] because we have the vision of what our society is supposed to be,” she said. “We should our responsibility seriously…and create timely and appropriate policies that will involve more stakeholders.”

Tshilidzi Marwala, vice-chancellor of the University of Johannesburg, cited reports – and his experience of – facial recognition software not recognising Africans. He added that the development of AI must “take into account…economic disparities [between nations] so that machines do not reflect the imbalances in the world”.

̽����Ƶ

Mikhail Strikhanov, rector of the National Research Nuclear University MEPhI in Russia, added that “society should not let [AI] develop without a plan”. Universities must work with policymakers to create new ethical frameworks, he added.

“Universities should be prepared to meet the challenge with faculty and research,” Dr Strikhanov said.

anna.mckie@timeshighereducation.com