Growing concerns that artificial intelligence could pose new cybersecurity threats

Whether you like the idea of ​​Artificial Intelligence or not, it is already part of your daily life. It helps you avoid traffic jams, recommends products to buy and powers our digital assistants.

But AI is also increasingly seen as a cybersecurity threat, capable of launching sophisticated attacks targeting banking, healthcare, infrastructure and elections.

“It’s used to trick people into thinking their child is being kidnapped and then handing them money to Venmo. And so these kinds of threats, what happens when you can’t trust someone’s voice information, are sort of an immediate threat to cybersecurity,” says Alexandra Givens, CEO of the Center for Democracy & Technology.

A recent CyberArk survey of 2,300 security professionals found that 93% expect threats from AI malware. And as AI capabilities increase, so does concern.

“What does this mean for, you know, the future ability to create chemical weapons or buy nuclear weapons? These are the types of concerns that are also being raised,” Givens says.

The Biden administration is addressing them, too, in a sweeping executive order issued last fall that calls for new safeguards for AI safety and security. Among other directives, it requires tech companies to share test results, federal agencies to set standards and calls for better protections for privacy, consumers and even civil rights. This is a first step that will also require action by Congress.

Last month, the United Nations passed the first-ever resolution to ensure AI is trustworthy.

“The risks and benefits of AI have the potential to impact us all. That’s why we all have to do it,” said Linda Thomas-Greenfield, Permanent Representative to the United Nations.

Some experts say one solution to countering AI threats is better AI.

“There is some hope that AI will actually improve cybersecurity defenses, because it could help us identify vulnerabilities, debug code and fix many of the flaws that adversaries exploit to carry out cyberattacks,” said Benjamin Boudreaux, policy researcher at RAND.

Ending AI-related cyberthreats will require regulation and accountability, experts say, not only from governments and private companies, but also from ordinary Americans, who will increasingly need to become fluent in AI.

“That means both understanding a little bit about how the technology works, but more importantly understanding its limitations and understanding that these technologies are very far from perfect,” Boudreaux said.