FAA Tests AI Air Traffic Tool as PwC Finds No Clear Owner for AI Risk

Phillip Buckendorf, co-founder and CEO of Air Space Intelligence, the Boston-based company building the software, said humans will still make key decisions. “I want to be very clear because this is important. This is not about automating air traffic control,” Buckendorf told NPR News. “What we're focused on is how do we optimize the airspace in its entirety.” He said air traffic controllers will remain responsible for keeping planes separated, while the FAA tool is aimed at traffic managers who oversee flight routing and scheduling. “The AI does the math, the AI does the computation,” Buckendorf said, but the human operator decides whether to accept recommendations. “They make the decision, they are the decider, they are the judge.”

The system, known as Strategic Management of Airspace, Routes and Trajectories, or SMART, draws on more than 200 data sources including weather, congestion, turbulence, winds and runway closures to identify more optimal flight paths, according to NPR News. Buckendorf compared it to a chess tool that can help traffic managers see several moves ahead. A similar ASI product, Flyways AI, is already used by Alaska Airlines and other carriers, but the SMART rollout will begin with a pilot project at three major airports in the Washington, D.C., area.

That plan drew criticism from Rep. Don Beyer (D-Va.), whose Northern Virginia district includes Ronald Reagan Washington National Airport. “The FAA should not use my constituents as guinea pigs for an unproven AI air traffic control system,” Beyer said in a statement. “Dangerously rushing a system that holds Americans' lives in its hands is not a good use case for artificial intelligence.” Buckendorf acknowledged fears about AI and air traffic control are real but encouraged critics to examine the technology. “The machine never makes a decision. It's a human operator, the air traffic manager,” he said. “We don't want them to run on 1980s software. Who can disagree with that?”

Dave Riley, a former FAA air traffic controller, called the SMART system “a step in the right direction,” but said it is not a “cure all” for larger problems such as a longstanding staffing shortage that forces many controllers to work mandatory overtime and grueling schedules. Riley said the FAA is overstating the software's impact on the flying public. “Throw the buzzword 'AI' on it and get everybody's attention,” he said. “That's what I see.”

The FAA's push comes as companies are still trying to determine who should be accountable when AI systems fail. ZDNet reported that PwC's Digital Trust Insights 2027 report, based on a survey of about 4,000 business and technology leaders across 71 countries, found no single role has clear responsibility for managing agentic AI or its security.

According to the survey, 29 percent of CEOs and security and risk leaders said accountability sits with the CIO, CTO or a similar technology role, while 17 percent said it lies with the CISO or cybersecurity teams. Twenty-six percent said responsibility should stay with a dedicated AI leader or function, and 11 percent said accountability is unclear because it is shared across multiple roles. About a third of organizations, 33 percent, have hired dedicated AI roles such as AI chief officers or AI board members. Forty-seven percent said cybersecurity is a standing agenda item for boards.

Jim Taylor, chief product and strategy officer at RSA, told ZDNet that agentic AI deployments need the same identity controls that have secured human users for decades. Each AI model or agent has an identity, credentials and varying access to resources, and can act on behalf of an employee. Taylor suggested organizations register sanctioned AI agents on a centralized platform, tie each agent to a human owner who must authorize high-risk actions, apply governance controls mapped to industry frameworks, and evaluate and decommission agents when they are no longer needed. “Companies will keep investing in AI, but they've brought on workers they don't see and can't control,” Taylor said.

Editor's Summary The FAA is testing a $875 million AI-assisted airspace tool that it says leaves human controllers and traffic managers in charge, even as critics question the pilot at Washington-area airports. A separate PwC survey found no consensus on who owns AI risk inside companies, while RSA says identity controls and human owners are needed for AI agents.