Many technology experts have warned that AI is headed for destruction, but a new study among top researchers has quantified the likelihood of AI causing human extinction.
A team of international scientists asked 2,778 AI experts about the future of the systems, with five percent reporting that the technology will lead to a collapse.
But a much scarier estimate came from one in ten researchers who said there is a shocking 25 percent chance that AI will destroy the human race.
The experts mentioned three possible causes: AI enables threat groups to create powerful tools such as artificial viruses, “Authoritarian rulers who use AI to control their populations and AI systems are exacerbating economic inequality by disproportionately benefiting certain individuals.”
Nearly half of researchers surveyed said there is a 10 percent chance that AI will destroy humanity
Controlling the regulation of artificial intelligence is the only answer to protecting humans, and if AI is not regulated, researchers estimate that by 2027 there is a 10 percent chance that machines will outperform humans in all tasks, but that chance could rise to 50 percent. by 2047.
The survey asked experts about four specific professions that would become fully automated, including truck drivers, surgeons, retail salespeople and AI researchers, and was told that there is a 50 percent chance that AI will be fully automated by the year 2116. will take over.
“It is an important signal that most AI researchers do not consider it very unlikely that advanced AI will destroy humanity,” said Katja Grace, one of the authors of the study. NewScientist.
“I think this general belief in a non-minuscule risk is much more telling than the exact percentage risk,” she added.
About 58 percent of researchers were more optimistic about AI adoption, saying there is only a five percent chance it will lead to human extinction.
The United Nations (UN) has already shared warnings about terrorist groups using AI for malicious intent, saying it will become yet another tool to add to their arsenal.
In the study, researchers were asked a series of questions to estimate how likely AI is to destroy humanity
'With a proven track record in the world of cybercrime, (AI) is a powerful tool that can potentially be used to promote or facilitate terrorism and violent extremism that fuels terrorism, for example by providing new modalities for physical attacks with drones or self-driving cars, increasing cyber-attacks on critical infrastructure, or enabling the spread of hate speech and incitement to violence in a faster and more efficient way,” UN Director Antonia Marie De Meo said in a UN conference . report.
This is because AI has contributed to 37 percent of workers being replaced by AI by 2023, according to a report by ResumeBuilder.
Hundreds of experts, including OpenAI CEO Sam Altman and Google DeepMind CEO Demis Hassabis, have made a one-sentence statement rack in May last year, demanding greater AI control and regulation.
“Reducing the risk of AI extinction should be a global priority, alongside other societal-scale risks such as pandemics and nuclear war,” the statement said.
Dan Hendrycks, the executive director of the Center for AI Safety, said in a separate letter rack at the time: 'As stated in the first sentence of the signatory page, there are many 'important and urgent risks of AI', not just the risk of extinction; for example, systemic bias, disinformation, malicious use, cyber attacks and weaponization.”
Even the 'Godfather of AI', Geoffrey Hinton, said he regretted creating AI, although he told The New York Times: 'I console myself with the normal excuse: if I hadn't done it, someone else would have done it.'
There is still significant uncertainty about the long-term effects of artificial intelligence, and the study acknowledged that “prediction is difficult, even for experts.”