What if terrorists get artificial intelligence?

An unmanned vehicle is the dream of many people, but it will become a nightmare if it falls into the hands of terrorists.

Rapid progress in the field is increasing the risk from those who are keen to take advantage of this technology to carry out the operations.

The prospect of unmanned cars crashing into the pedestrian crowd, government buildings attacked by drone equipment may soon become a reality in the next 5-10 years if the agency No responsibility to prevent.

The world is concerned about artificial intelligence (AI) turning against humanity.But the closest scenario is the person who uses AI to resist homosexuality - (Photo: GETTY).

But AI applications are not as simple as that. What if more sophisticated AI applications fall into the hands of those with bad intentions?

It is also the question that 26 leading scientists from Cambridge, Oxford, UK and Yale universities in the United States, together with military experts, responded to in a recent study.

The work announced on February 21 triggered a warning bell about the risk of criminals and "lone wolf" attackers using AI.

"AI is a game changer. What will the world be like in this study in the next 5-10 years , " Dr. Seán Ó hÉigeartaigh of Cambride University told the British Independent newspaper.

The researchers identified that the use of AI with malicious intent poses a threat not only to cyberspace, but also to other political consequences. More dangerous, with AI, terrorist attacks will become more accurate, less time consuming, easy to hide but most serious - damage at an unprecedented level.

The 98-page study has warned that the cost of conducting AI attacks is lower than that of human attacks.

With the power to synthesize images, texts and sounds to imitate and impersonate others, AI is powerful enough to influence and control public opinion. This document also calls on governments and technical experts and policy makers to come up with solutions to prevent these hazards.

The Singapore Air Force Heron 1 unmanned aircraft is displayed at the Singapore Airshow on February 8.With scientific progress, unmanned aircraft now look "terrible" like real fighters - (Photo: REUTERS).

"There are things we have to do right now , " stressed Dr. Seán Ó hèigeartaigh, "a more comprehensive, more comprehensive approach is needed. For example, the design of software and hardware that is hard to attack hackers. It is also important to discuss what part of national law or international norms for AI is to discuss. "

The report also questions whether scholars and experts should limit disclosure or disclosure of new developments in the field of AI when there are still no measures to respond to potential threats. the AI ​​created.

AI technology includes the use of computers to perform tasks that often require human intelligence, such as making decisions or identifying texts, words and images.

Despite many benefits, AI development is also warned to bring potential risks. The billionaire and famous astrophysicist also repeatedly warned about the downside of developing and using AI.