A survey of 2,700 AI researchers shows high potential of AI causing catastrophic outcomes such as human extinction.

Concerns raised include deepfakes, public opinion manipulation, engineered weapons, authoritarian control, and economic inequality.

Most respondents believe a 'superintelligence' could have a 5% or greater chance of causing an extinction-level event.

The survey underlines the need for ethical and safety considerations in AI development.

Humans face challenges in predicting the consequences of new technologies.

Lawmakers in Europe and the United States are developing regulations to address potential risks associated with AI.