David Dalrymple, a well-known AI safety expert, made an important warning in an interview with The Guardian.
Elchi.az reports that Dalrymple, director of the UK’s Advanced Research and Scientific Discovery Agency (ARIA), stressed that the world may not have enough time to prepare for the current risks:
We should be concerned about systems that can completely replace what people do to perform tasks on Earth. We will be defeated in all areas where we must dominate to manage civilization, society and the planet.
Dalrymple stated that there is a significant gap in understanding between governments and AI companies regarding upcoming technological advances:
Developments are happening very quickly and we may not have enough time to prevent them from a security point of view. The idea that machines will perform the most economically valuable tasks better than humans in 5 years is not something out of science fiction.
Dalrymple noted that within the framework of ARIA, which is funded by the state but operates independently of the government, they are developing systems to control the use of artificial intelligence in important areas such as energy.
Touching on artificial intelligence, he said: “We cannot assume that these are reliable. Given the economic pressures, it is unlikely that we will have the science in time to develop them safely. Therefore, the second best thing we can do is to control and reduce their negative aspects.”
Mustafa Suleyman, CEO of Microsoft’s AI company, also warned that if appropriate regulations are not implemented, the developing AI technology may become “unmanageable” in the next few years.
Şayəstə