AT&T CEO John Stankey has urged critical infrastructure operators to coordinate faster on AI security risks.
Speaking at Semafor’s Silicon Valley and the World conference, Stankey said AI changes the threat landscape for every major industry. Telecoms, energy, finance, transport, water, and IT now face faster attacks.
“At this moment in time, I think everybody’s more vulnerable,” said Stankey.
However, he did not present a bleak view of AI’s future. Instead, he argued that the technology can improve infrastructure security over time. AI can help detect weak points, analyze incidents, and speed response.
The challenge is timing. Attackers can also use AI to move faster. They can scan systems, test vulnerabilities, and scale campaigns with less effort. The UK National Cyber Security Centre recently described AI as a “force multiplier” for cyber threats.
Stankey said industries need better information sharing. He placed more emphasis on coordination than regulation. In his view, companies must work directly with each other. They must also engage AI developers and public agencies.
“The meat’s going to be put on the bones by businesses and industry working with each other, working with those that are generating the new technology … And it’s going to require better sharing,” he said.
That approach could move faster than new rules. Telecom operators already exchange data on fraud, outages, and cyber events. Yet AI creates a wider problem. A threat found in one sector may quickly affect another.
For that reason, Stankey highlighted the Alliance for Critical Infrastructure. The group launched earlier this year. It includes more than 50 companies across communications, energy, finance, IT, transport, and water systems.
Verizon CEO Dan Schulman sits on its board with Stankey. Other telecom members include Charter Communications, Comcast, Lumen Technologies, and T-Mobile.
Stankey also pointed to Anthropic and its Project Glasswing rollout. He described it as one useful model for controlled testing. The effort examines security risks before powerful models reach wider use.
Still, this method creates difficult decisions. Companies may identify major issues before they can fix them. Stankey warned that some fixes could take two years or more.
“Now there’s a long list of things that have to be addressed and fixed, many of which are going to stretch out for two years or more. The question, how much risk do you want to take during that period, and how do you mitigate that?” he said.
For telecom networks, the message is clear. AI will not wait for perfect governance. Operators must improve trust, share intelligence, and act before risks spread.

