Kári Stef podcast explores AI’s promise and risks

Artificial intelligence could bring significant opportunities to Icelandic society, but its risks must also be considered as the technology grows more powerful, mbl.is reports. Those issues are discussed in a new podcast hosted by Kári Stefánsson.

The guests are Hafsteinn Einarsson, dean of industrial engineering, mechanical engineering and computer science at the University of Iceland; Stefanía Benónísdóttir, a research lecturer and academic lead of the university’s AI centre; University of Iceland professor Magnús Örn Úlfarsson; and Jóhann Páll Jóhannsson, the environment, energy and climate minister.

The discussion ranges from how AI works and how quickly it is advancing to its possible effects on education, the labour market, science, public administration and democratic debate. The guests also consider warnings from people in the AI industry and calls to slow development.

One scenario raised is that highly advanced AI could become uncontrollable and threaten humanity’s existence. Einarsson says he does not see an extinction risk from AI as it exists today, while stressing that this view does not address where future developments might lead.

Magnús Örn says AI’s current capabilities may be overestimated, noting that it has not cured diseases or solved environmental and climate problems. He remains concerned, however, that the technology could get out of control or be used maliciously, such as to break into bank computer systems. He says regulation is needed, while warning that it could also prevent AI from being used to solve real problems.

Kári says he leans towards believing AI’s benefits outweigh the risks likely to emerge in the coming years, though he adds that this may reflect too much faith in humanity.