Shafaqna English- Warnings that advanced AI could eventually threaten human survival are intensifying, with some experts putting extreme scenarios on the table—while others argue that near-term risks such as misinformation, job disruption and weak safeguards deserve far more attention, according to The Guardian.
The debate intensified this week in the UK and US. Former UK defence secretary Des Browne and computer scientist Stuart Russell compared the potential threat from artificial superintelligence (ASI) to nuclear weapons, while Anthropic alignment researcher Evan Hubinger said he believed there was a greater than 10% chance AI could kill all humans within the next decade.
The warnings come as AI companies compete to develop increasingly capable systems. Former OpenAI researcher Jacob Coxon resigned while criticizing what he described as a race among leading labs to reach advanced AI first. UK officials have also raised questions about access to the latest models for independent safety testing.
Some policymakers are calling for international action. UK MP Darren Jones has proposed a multinational treaty for the safe development of superintelligence, while US Senator Bernie Sanders is urging Congress to strengthen AI regulation.
But not all experts accept the most catastrophic predictions. Researchers including Andrew Rogoyski, David Barber and Sandra Wachter argue that current AI systems remain far less capable and versatile than humans and that focusing on “Terminator” scenarios can distract from immediate problems.
Those risks are already tangible: AI can amplify misinformation, disrupt employment, consume significant resources and introduce vulnerabilities when connected to critical systems. The challenge, experts broadly agree, is finding a way to capture AI’s benefits while imposing meaningful safeguards on increasingly powerful systems.
The key question is no longer simply whether AI will become more capable—but whether governments and developers can build effective safety and coordination mechanisms before capability outpaces control.
Source: The Guardian