Agentic AI is supporting cyber criminals, helping them discover vulnerabilities in networks even faster than before. That’s why it’s critical for trucking companies to invest in AI to protect themselves from the bad actors using AI against them, said Joe Ohr, chief operating officer at the National Motor Freight Traffic Association (NMFTA).
“Because they're using AI, you have to use AI to keep your head above water,” Ohr said.
Some have already figured that out.
Ohr said he knows of one company using AI to detect anomalies in their network. For example, if someone logs into their network from an unusual location or at an irregular time. AI can detect patterns and find anomalies. Using AI, this company found an anomaly in its network and was able to shut it down within five minutes, he said.
[RELATED: How to keep your business running when a cyberattack occurs]
“Sure enough, about 15 minutes later, they found out they had a bad actor in the network, and they got ransomwared, but because they shut it down so fast, no harm, no foul,” Ohr said. “If it wasn't for the AI detecting that, they would have been in trouble” because they were informed by other vendors and the government well after the fact.
The attack landscape is getting more complicated and crazier by the day, said Perygee Co-founder and CEO Mollie Breen, and AI is part of that. Every single cyber product has AI in it at this point with a goal to increase security posture, she said.
That could look like identity and access control or a firewall setting that ensures certain traffic is blocked. It could also be a vulnerability scanner like Ohr mentioned that highlights vulnerabilities a security team wasn’t previously able to find. Breen said AI is enabling organizations to level up their security teams.
She said the current focus of AI in cybersecurity is for threat and anomaly detection, such as identifying phishing emails and suspicious IoT devices.
Prior to AI, Ohr said a cybercriminal would find their way into a network and remain there for six or more months to gather information before they were able to do damage.
Now, that damage is done in minutes.
A manual process of detecting a cybercriminal could take days to weeks using a natural threat hunt system in which human analysts search for threats within a network, said Artie Crawford, director of cybersecurity at NMFTA.
He said using AI to defend against bad actors using AI could make the difference in blocking an attack within minutes versus days or weeks.
“We've gone from looking for a needle in a haystack to using agentic AI to look for a needle in a pile of needles,” Crawford said. “AI is allowing the bad actors to generate these attacks quicker, especially on the asset side. But also on the defensive side, those anomalies are able to be detected equally as fast.
“The question is, where do we put a human in the loop to ensure what the response should be,” he added.
With AI poised to take over repetitive tasks and even potentially replace some employees, it’s possible that the AI could eliminate human vulnerability and improve cybersecurity risk.
Crawford said social engineering – like phishing, smishing and vishing, among others – is still probably the No. 1 entry point into any organization, making humans the most susceptible to cyberattack attempts.
AI can allow for faster detection, blocking and response, but he said there still always needs to be a human in the loop to adjudicate what the AI is set to do.
“When we start allowing AI to set and redefine the rules, then we start to get into this period of time where AI is going to start to self-learn,” Crawford said. “We have to ensure that the boundaries and the ethical use of AI is proper. So there needs to be a human in the loop no matter what, especially when it comes to AI allowing itself to learn and become better at what it does.”
