Webinar replay: Cybersecurity Challenges, Brought to You by AI

The advent of AI in manufacturing has brought about many efficiency gains and benefits such as better ROI for their operations, which includes increased uptime performance and efficiency.

But AI also has opened manufacturing systems (especially OT that wasn’t made to be networked but now is in many places) to a much larger threat landscape.

For a special webinar before Labor Day that was part of our Digital Transformation Academy—a partnership between Smart Industry and our sister brand IndustryWeek and sponsor Integris—SI enlisted two subject-matter experts who specialize in manufacturing cybersecurity to talk about the special challenges that AI brings to modernized plant systems and environments.

Tim Freestone is chief strategy officer and AI-first strategist with Kiteworks, vendor of a software platform designed to manage and secure sensitive and proprietary data communications.

Carolyn Duby is field CTO and cybersecurity lead at Cloudera, vendor of a data platform designed for flexible data management and analytics across any cloud and on-premises environments.

Freestone and Duby joined Smart Industry’s Scott Achelpohl for the hourlong conversation, which was slowed by some technical problems with Scott’s video but was nonetheless spirited. Register (if you haven’t already) and watch the replay anytime your schedule allows.

The risk ... now that AI is getting involved

Scott added some valuable data points, courtesy of Rockwell Automation and Black Kite:

Cyberattacks are No. 2 in terms of external risk to growth.

As a result of the cyber threats, one-third of manufacturers plan to strengthen IT/OT architecture security in the next five years.
Almost half of manufacturers, 49%, plan to use AI or machine learning for cybersecurity (up from 40% in 2024).
38% are leveraging data for cybersecurity protection (up from 31% last year).

More came from Tim and Kiteworks’ just-released AI Data Security and Compliance report:

83% of organizations lack basic AI security controls.
Only 17% of the surveyed organizations have technical safeguards.
26% admit that more than 30% of data sent through AI tools contains private information—trade secrets, designs, and customer data. The industry relies dangerously on employee training over technology for its cybersecurity.

Tim added that the AI security gap in manufacturing has created the “perfect storm” for cyber risk.

e-handbook: Cybersecurity

While manufacturers are ahead of other industries in implementing hard controls (27% block public AI versus 17% overall), he contributed, the convergence of three factors creates unprecedented risk:

83% of organizations still lack basic AI security controls.
Manufacturing remains the No. 1 ransomware target at 21% of all attacks.
Nearly half of manufacturers (49%) are rushing to adopt AI for cybersecurity itself.

This rapid adoption without proper safeguards, he said, is like “building a fortress with an open drawbridge”—especially concerning when 26% admit that over 30% of data flowing through AI tools contains critical IP like trade secrets and designs.

Carolyn noted that manufacturing is adopting AI to drive efficiency, which also means more systems—especially OT—are being connected in ways they weren’t originally designed for.

She added in materials prior to the webinar: “The big issue with cyber data is that analytics and AI are essentially add-ons. The data is not ready for AI and there is not really a vendor-agnostic proper AI platform for cyber data. There are logging appliances, but they only do what they are programmed to do.”

She also noted that OT systems were designed for uptime and performance—not for connectivity or cybersecurity. As manufacturers start layering in AI and cloud technologies, they’re exposing systems that were never intended to be networked. That makes them attractive to attackers looking for easy entry points and maximum disruption.