Artificial Intelligence has become a serious matter of concern in the field of science and technology.It has the potential of rising up and transforming our society, but there’s also a possibility of annihilation where AI may prove dangerous to the human race, as illustrated in the
It has the potential of rising up and transforming our society, but there’s also a possibility of annihilation where AI may prove dangerous to the human race, as illustrated in the Terminator movie sequel.
The latter possibility is exactly what researchers from Google and the University of Oxford are trying to mitigate. A research paper published by the Future of Humanity Institute mentions how an emergency countermeasure taken by humans can be realized in a way to mislead the Artificial Intelligence.
In the presented approach the computerized systems are made “believe” that the actions were actually its own decisions.
The researchers from Google and Oxford explain how using a “big red button”-approach with different instructions could make AI operate under the impression that interruption is external and will not repeat itself. This way, it would not assume future interruptions and may never learn to protect itself against the actions of human controllers.
Along Google and Oxford, many notable names already acknowledge how big a threat AI can be. Recently, Elon Musk, the Tesla, and SpaceX CEO, donated $11 million to the Future of Life Institute for further research on AI safety.
You can know more about the working of these algorithms by checking the original paper here.