Concerns for an AI apocalypse rise in last year

Matthew SmithHead of Data Journalism
June 05, 2023, 1:07 PM GMT+0

There has been a 10pt increase in the number who see artificial intelligence as a top threat to human survival

2023 has been the year of artificial intelligence, with the maturation of programs like ChatGPT capturing the public and media imagination.

At the same time, increasing sophistication of AI has led to warnings that the technology could lead to the extinction of humanity, with a statement by the Center for AI safety calling for safeguarding to be given as much priority as pandemics and nuclear war endorsed by senior figures working in the field.

With such sensational claims making newspaper headlines, it is no surprise that a new YouGov survey shows that public concerns about the threat of artificial intelligence to the human race have suddenly increased after six years of tracking.

In 2016, asked to choose up to three most likely causes for human extinction, the number of Britons choosing ‘robots/artificial intelligence’ was just 5%. As recently as February 2022 that figure was still only 7%. However, since then it has risen by 10 points to 17%, the only category to see an increase in the last 15 months.

Nevertheless, this still leaves artificial intelligence far behind first-placed nuclear war (55%), as well as climate change (38%) and pandemics (29%). And the public are no more likely than they were last year to believe a human extinction event is likely during their lifetime (9-11%).

Most Britons back contingency plans for AI

In response to the warnings from the Center for AI Safety, Rishi Sunak said last week: “People will be concerned by the reports that AI poses an existential risk like pandemics or nuclear wars – I want them to be reassured that the government is looking very carefully at this.”

Most Britons are likely to approve of this, as 64% say that the government should be developing realistic contingency plans to deal with threats from robots and artificial intelligence. This is a 20 point increase since last year, and a 37 point increase since 2016.

This puts AI on the same contingency planning level as meteorites (62%) and extreme seismic events like volcanic eruptions or tsunamis (64%), but still far behind the desire to prepare for nuclear threats (88%), pandemics (87%) and climate change (82%).

While Britons will appreciate contingency planning, there is a notable lack of confidence in the government to manage the emergence of AI in day to day society: a separate YouGov study found 68% have little to no confidence in the government to effectively regulate the development and use of AI.

While it is clear that many Britons consider artificial intelligence to have potential as a threat, that is not their dominant impression of the technology. The same YouGov survey recently found that only 17% of Britons’ first thoughts on the topic of AI were to do with its apocalyptic potential, with the public far more likely to say instead that their minds go to its everyday implications, like the effect it might have on jobs and society (52%).

See the full results here

Photo: Tri-Star Pictures