AI poses a major threat to companies, according to Allianz

In the new "Risk Barometer" published by corporate insurer Allianz Commercial, AI has shot up from tenth to second place behind the long-standing leader, cybercrime.

The two are linked: criminal hackers are increasingly using AI for their attacks. However, according to the risk barometer, the use of AI can also be dangerous for a company without any malicious intent, for example if managers and employees make the wrong decisions based on incorrect data and information.

The three main risks are interrelated

Business interruption is the third biggest global business risk this year. This is also linked to cyber attacks: A common cause of business interruption is online extortion: hackers paralyze a company's computer systems via encryption and demand large sums of money for the subsequent decryption.

Allianz Commercial is a subsidiary of the Munich-based Dax group; the company publishes its "Risk Barometer" every year at the beginning of the year. The assessments are based on a survey of 3338 experts from 97 countries last fall. They include executives and managers from other companies, risk and claims consultants, insurance brokers, experts from industry associations as well as Allianz employees.

The answers of the respondents differed from country to country, but not fundamentally: AI risks were ranked second in Switzerland, fourth in Germany and first in Austria.

AI is both a curse and a blessing

AI is therefore a double-edged sword: a majority of companies see the technology as an opportunity, not least for the automated defense against malicious cyberattacks. At the same time, however, many of the experts surveyed see major dangers: AI harbors an ever-accelerating risk, said Michael Furtscheller, the regional business manager for Germany and Switzerland - "perhaps also a curse and a blessing".

According to him, cyber criminals use AI to perfect "social engineering", among other things, to pose as managers and deceive their subordinates. "By writing highly tailored emails that they want you to click or do something else, whether it's cloning or generating speech, or even faking videos," explained Michael Daum, head of cyber claims handling. "The vast majority of attacks we see still require the involvement of a human - usually an employee - to enable the attack."

AI poses a double threat to companies from outside and inside

However, according to the Allianz managers, external attacks are only one side of the problem. The completely legal use of AI software by the company's own employees and managers also poses risks for companies. "By definition, AI works with a certain degree of autonomy, which is why the results can be incorrect or fictitious," said Allianz Commercial Manager Alexandra Braun.

"And incorrect or even discriminatory AI results can of course also lead to legal disputes or negative press coverage and then to a loss of reputation and image for companies." AI's own risks also include copyright infringements if the software copies or uses protected information.

Wide range of other risks

The other risks in the global top ten range from politics and nature to the traditional uncertainties of business life. In fourth place is legislation and regulation, which refers to both US customs policy and other trade barriers as well as the bureaucracy that is lamented in many countries.

The next two places are taken by natural disasters and climate change, followed by political instability and violence, negative economic developments such as inflation, fire and explosions. Uncertainty about market developments, be it new competitors, company takeovers or other changes, takes tenth place.