JPMorgan Chase CEO Jamie Dimon is personally recruiting some of America’s largest companies for a new push to confront the risks artificial intelligence could pose to banks, energy grids, water systems, telecommunications networks and other critical infrastructure.
The outreach has reached more than 40 companies across financial services, energy, water, utilities, telecommunications, airlines, railroads and other technology-dependent industries, according to people familiar with the effort. Dimon has contacted CEOs at major and regional banks and technology companies as part of the campaign.
“JPMorgan Chase CEO Jamie Dimon is urging corporate leaders to join a U.S.-focused industry group to address risks posed by AI, as corporate America rapidly adopts the developing technology,” Reuters reported, citing two sources familiar with the matter.
The effort is being built around the Alliance for Critical Infrastructure, or ACI, an industry group JPMorgan helped establish with Mastercard, Berkshire Hathaway Energy and other companies. The goal now is to broaden the group’s focus as AI becomes embedded deeper into the systems that keep the U.S. economy running.
That makes this more than another corporate AI working group.
The companies Dimon is trying to bring together operate infrastructure where a software failure, cyberattack or compromised AI system could have consequences far beyond a corporate network. Banks move money. Utilities keep electricity flowing. Telecommunications companies carry communications. Airlines and railroads move people and goods. Water operators provide a resource millions of Americans depend on every day.
AI is entering all of those systems at the same time.
Dimon is taking his AI warnings directly to corporate America
The outreach began in July, with ACI and Dimon contacting prospective members and seeking to schedule discussions in August. The organization has yet to disclose how many companies have formally joined the effort.
The plan is to develop a common picture of how companies are deploying AI, what new risks are emerging and what safeguards may be needed. Participants are expected to share information across industries and work with the Trump administration on technology and AI-related threats.
The revamped ACI is expected to be fully operational by the end of the year, according to a person familiar with the plans.
For Dimon, the effort puts organizational muscle behind warnings he has been making publicly about increasingly capable AI systems.
His voice carries unusual weight in corporate America. JPMorgan is the largest U.S. bank, putting Dimon at the intersection of finance, technology, cybersecurity and government policy. He has spent years discussing AI’s potential to transform businesses, yet he has become increasingly vocal about what happens when advanced systems fall into the wrong hands.
According to Reuters, Dimon said ACI leaders recognized the issue before the current wave of concern.
“We are proud to support this important work,” Dimon said, adding that ACI leadership had seen the need to prioritize AI “years ago and got critical infrastructure companies working together.”
The AI risk is moving beyond chatbots
Much of the public conversation around artificial intelligence has centered on productivity, jobs, coding, search and the race among companies building increasingly capable models.
Critical infrastructure raises a different set of questions.
AI models can help companies detect threats, automate operations, analyze enormous amounts of information and identify weaknesses humans might miss. Those same capabilities can potentially help attackers discover vulnerabilities, automate cyber operations or make sophisticated tools accessible to people who previously lacked the technical expertise to use them.
Recent cyberattacks against water systems in Minnesota and other states have added urgency to the conversation around information sharing between industries.
ACI said the challenge requires companies and government to work together.
“At a time of growing cyber threats, protecting the systems Americans rely on every day requires strong collaboration between government and critical infrastructure companies,” the organization said.
The group’s proposed role resembles an early-warning network across industries. A vulnerability discovered by one company or sector could carry lessons for organizations elsewhere rather than remaining trapped inside a corporate security department.
That cross-sector structure could become one of the initiative’s most significant features.
A bank and an electric utility may operate very different businesses, yet both increasingly depend on cloud infrastructure, software vendors, telecommunications networks, identity systems and AI models. A weakness somewhere in that technology chain can travel much further than the company where it was first discovered.
Dimon’s “ballistic missiles” warning
Dimon has used unusually stark language to describe what could happen if highly capable AI systems become widely accessible without adequate controls.
In July, he warned about Anthropic’s Mythos AI model and the potential consequences of giving individuals access to advanced capabilities.
“You’re giving ballistic missiles to individuals with Mythos,” Dimon said, according to Reuters.
His concern centers on a question AI companies, governments and corporate security teams are increasingly confronting: what happens when models become good enough to help people perform tasks that once required specialized expertise?
That question becomes far more serious when the targets are financial institutions, energy networks or public infrastructure.
Banks are separately testing Mythos through another industry initiative. The ACI project is distinct from that work and has a much broader cross-industry mission.
Washington is building its own AI security network
Dimon is not alone. Corporate America’s effort is emerging alongside a similar push from the federal government.
In July, the U.S. government launched the Gold Eagle initiative, bringing AI developers, critical infrastructure operators and federal agencies together to exchange information about vulnerabilities uncovered by advanced AI models and coordinate responses.
The timing is significant.
AI security has often been discussed as a problem for the companies developing the models. The newer approach treats it as a problem shared by model developers, companies operating critical infrastructure, and government agencies responsible for national security.
That creates several layers of defense.
AI companies can identify dangerous capabilities inside their models. Infrastructure operators can determine whether those capabilities expose weaknesses in real systems. Government agencies can help coordinate responses when risks cross companies, industries, or national borders.
ACI could provide another layer by connecting corporate operators directly with one another.
Why 40+ companies matter
The scale of Dimon’s outreach may be the most revealing part of the story.
This isn’t JPMorgan assembling a handful of banks to discuss financial cybersecurity. More than 40 companies spanning industries that rarely sit around the same table have been approached.
That suggests corporate leaders are beginning to view advanced AI risk less as an IT department issue and more as a shared infrastructure problem.
It marks a shift in how the AI boom is entering its next phase.
The first corporate race centered on adoption: Who could deploy AI fastest? Who could automate more work? Who could build better models? Who could turn AI into new products or lower operating costs?
A second race is forming around resilience.
Companies now have to figure out what happens when AI systems fail, behave unpredictably, expose sensitive information or give attackers capabilities that previously belonged to highly skilled security teams.
The companies responsible for America’s financial networks, electricity, communications, transportation and water infrastructure have little room for experimentation when failures can affect millions of people.
That is the problem Dimon’s coalition is trying to get ahead of.
From AI adoption to AI preparedness
There is an interesting contradiction at the center of corporate America’s AI boom.
Companies are investing heavily to put increasingly capable AI systems inside their businesses. At the same time, some of the executives leading that adoption are becoming louder about the risks those systems may create.
Dimon’s campaign shows those two positions can exist together.
The question is no longer whether corporations will use AI. That decision has largely been made. The harder question is how companies can deploy it without creating vulnerabilities that stretch across interconnected industries.
ACI’s success will depend on whether companies are willing to share meaningful information about weaknesses, incidents and emerging threats, including information they might normally prefer to keep private.
Getting 40 companies into a room is one thing. Getting banks, utilities, technology companies, airlines, telecommunications providers and infrastructure operators to openly exchange sensitive security information is much harder.
Yet that may be exactly what advanced AI requires.
Cybersecurity threats do not respect corporate boundaries. AI could make those boundaries matter even less.
Dimon’s bet appears to be that corporate America will be safer if the companies running America’s critical systems start treating AI risk as a shared problem before a major crisis forces them to.
