AI could make bugs and cyber attacks more devastating - here are the grim predictions of a risk expert

AI could make bugs and cyber attacks more devastating - here are the grim predictions of a risk expert

  • In a study, the insurer Allianz details the risks that could arise with the development of artificial intelligence and robotization.
  • The fact that our society is becoming more and more connected makes the consequences of a hypothetical global cyber attack more and more dramatic.
  • Allianz goes so far as to imagine what might happen if AI-controlled autonomous weapons are developed.

Will artificial intelligence lead the world to a Terminator-style scenario?

In a study entitled “The Rise of Artificial Intelligence: Future Outlook and Emerging Risks”, the Allianz Insurance Group reviews the various benefits and risks that could emerge with the development of AI.

In particular, the insurer’s document depicts a worrying future in the event of ill-intentioned use of artificial intelligence in an increasingly connected society.

Here’s how artificial intelligence could be used by ill-intentioned people:

Hackers will be able to use AI to launch sophisticated cyberattacks:

“Misuse of a strong AI can also increase the risk of cyber attacks if malicious hackers train this AI to attack.”

The new machines that populate the ground and air can be hijacked and become connected weapons:

“Autonomous weapons, such as drones, could also be used. This type of risk is often underestimated.”

A disaster scenario that alarms researchers in AI and robotics. More than 3000 of them — as well as personalities such as Stephen Hawking, Elon Musk or Steve Wozniak — signed, in 2015, an open letter, to alert about the danger of autonomous weapons that would make the decision to kill without human intervention.

Hackers will also be able to develop these tools faster thanks to AI, Allianz believes:

“AI could facilitate more serious incidents by lowering the cost of new tools and weapons that would allow attacks to be launched.”

And the scale of the problems will be magnified.

“Increased inter-connectivity means that the vulnerability of automatic, autonomous or self-learning machines to bugs and cyber-attacks will only increase,” observes Allianz.

If an IT or electrical infrastructure is affected — by a bug or a cyber attack — other equipment connected to it may also be affected.

Moreover, since with machine learning, ais have the ability to learn from each other, “a programming error or a hacker attack could be replicated on many machines”. An affected machine will then be able to “repeat the same error several times, which will cause an unexpected accumulation of losses and difficulty in identifying the source of the problem,” Allianz warns.

With the result, major economic losses:

“Whatever the cause, an unintentional error caused by AI could turn into a major accident, which could damage a reputation and therefore financial results,” the study believes.

In 2017, Cyence estimated the economic losses caused by the WannaCry virus, which reportedly affected hundreds of thousands of computers in 150 countries, at $4 billion. For the insurance company Llyod’s, a major global cyberattack could cost 53 billion dollars, as much as Hurricane Sandy that hit the United States in 2012.

Receive our latest news

Every day, the main Business Insider news

Go to our cases Get a free quote