Amid a flurry of Google and Microsoft generative AI releases last week during SXSW, Garry Kasparov, who is a chess grandmaster, Avast Ambassador and Chairman of the Human Rights Foundation, told me he is less concerned about ChatGPT hacking into home appliances than he is about users being duped by bad actors.
“People still have the monopoly on evil,” he warned, standing firm on thoughts he shared with me in 2018. Widely considered one of the greatest chess players of all time, Kasparov gained mythic status in the 1990s as world champion when he beat, and then was defeated by IBM’s Deep Blue supercomputer. That moment came to represent an awakening that machines might one day overpower humanity.
“AI wins not because it’s smarter than humans, but because it simply makes less mistakes,” he explained. “I never understood this fear of AI – it’s a useful tool that does what we tell it to do. I’m a big proponent of human-machine collaboration.”
What keeps him up at night is far more sinister, he said. In his role with the cybersecurity firm Avast, which was acquired in 2022 by NortonLifeLock Inc., and is now part of the Gen Digital family of products, the company found that while 30% of cyberattacks on its platform targets vulnerable networks with weak security, about 70% are phishing and other scams in which users are willingly giving their information to bad actors.
“These cognitive threats are a game changer and not something a firewall can be built against,” Gen Chief Technology Officer Michal Pechoucek told me. Social engineering attacks can be particularly effective when someone impersonates a close contact in your inner circle like your boss and instructs you to do something, like pay an invoice. With generative AI tools, it’s getting easier for scammers to write more convincing text because reinforcement learning is helping the algorithm fine-tune messaging to optimize the clickthrough rate. With all of the details of our day-to-day lives being shared across social networks, it’s also allowing for more nefarious personalization.
“It comes down to the consumer making the right decision,” Pechoucek said. Simple things people can do to protect themselves include checking the sender’s email to ensure the address matches the identity, not clicking on links or downloading PDFs, never reusing passwords, and sharing less on social media, he advised.
“It’s a lazy criminal going after a lazy customer,” Kasparov said. “Step up your defenses a little bit and they’ll move on.”
With regards to generative AI hacking into systems, Pechoucek said he was aware of a study by cyberthreat intelligence agency Check Point Research that showed ChatGPT being tricked into creating malware, going against its own rules. However, he said Gen uses generative AI in its own large language models to ensure its defense is strong enough to catch whatever new viruses criminals come up with.
Pechoucek says he can see a situation where people start to question the very reality of everything, especially with the advent of text-to-video deepfakes. One solution might be a platform where users can verify media with encrypted metadata, this could build digital safety and trust, he said.
Regardless of how innovation in the space plays out, Kasparov who ran as an opposition candidate to Russian president Vladimir Putin and left in exile a decade ago, said the US doesn’t have much to fear with regards to global cyber warfare with state-controlled adversaries. Assuredly he said, “Free people in free markets will always have the upper hand.”