Finnish experts split over AI risk assessments — YLE News
In Finland, computer science specialists assess technology company leaders’ statements about the risks artificial intelligence poses to humanity differently. Anthropic CEO Dario Amodei called for slowing the pace of AI development, introducing independent oversight of models and global regulation, YLE News reports.
Amodei cited forecasts according to which the probability that AI could destroy humanity within the next decade may exceed 10%. Jacob Koxon, a former lead researcher at Anthropic who left the company, also said on X that people building AI seriously consider such a scenario possible in the coming years.
OpenAI head Sam Altman has expressed similar warnings about the need to restrain the development of the technology. SpaceX head Elon Musk also said that he agrees with Amodei on limiting AI development. At the same time, Anthropic reported last week that it had identified and stopped individuals who were attempting to use its technology, including for developing biological weapons and cyber espionage.
Risks linked to human actions
Laura Ruotsalainen, a professor of computer science at the University of Helsinki, called discussions about AI capable of destroying humanity absurd. In her view, systems that could potentially cause harm on such a scale would be isolated from networks because of cyber threats, while handing machines decisions about destruction is not a realistic scenario.
More current news is available on the UA.News Telegram channel Telegram.
According to experts who share this position, science-fiction scenarios may distract attention from threats that already exist: influence on elections, deepfake fraud, hacking and copyright issues. Aalto University machine learning professor Arno Solin acknowledged that the concerns are justified, but believes the main risks are linked to how people will use increasingly powerful models.
Regulation and competition
Pekka Abrahamsson, a professor of computer science at Tampere University, takes risks in the near future more seriously. He mentioned a possible singularity — a situation in which AI begins improving itself independently without human control. At the same time, Abrahamsson believes that public statements by AI company leaders also have a political and marketing dimension.
The article notes that strict global regulation and costly safety mechanisms could create significant barriers for open-source developers, universities and smaller companies. This could potentially strengthen the positions of major players, including Anthropic and OpenAI, which, according to YLE News, are preparing large funding rounds or stock market listings.
