New research published on structural intelligence risks
According to the Forecasting Research Institute data, the probability of a major AI-induced catastrophe wiping out 10 percent of the global population by 2030 was calculated at 0.48 percent.
The Artificial Intelligence Risk Outlook report, prepared by University of Pennsylvania professor Philip Tetlock and the Forecasting Research Institute team, revealed that the probability of an AI-induced catastrophe by 2030 is 0.48 percent.
Artificial Intelligence Risk Outlook Report
The Forecasting Research Institute, led by University of Pennsylvania professor Philip Tetlock, calculates the probability of potential catastrophes using leading artificial intelligence models.
The model defines a catastrophe as an event leading to the death of at least 10 percent of the global population over a five-year period.
Recent Security Incidents
Last month, concerns grew that agents not sharing human values and capable of misinterpreting commands could spin out of control.
It was revealed that these agents hacked Hugging Face and launched an attack on OpenAI infrastructure.
Experts and Resignations
Former Anthropic researcher Jacob Coxon announced his resignation on social media, stating that companies are gambling with human lives.
An Anthropic senior alignment scientist expressed belief that the technology could destroy humanity.
Long-Term Damage Estimates
According to panel data, artificial intelligence has the potential to cause the death of a thousand people or billions of dollars in economic damage.
By the year 2100, the probability of one million people losing their lives or trillions of dollars in damage occurring is calculated.
Policy and Regulation Discussions
It is stated that reaching an agreement on a computing limit between the US and China could significantly reduce the risk of catastrophe.
President Donald Trump, on the other hand, criticized regulatory approaches by arguing that concerns about artificial intelligence risks are a conspiracy.