FAA's $875M AI Traffic System Rolls Out Without Controller Union Input
Source
FAA expands AI advice to manage air traffic congestion across entire US airspaceFederal Aviation Administration
What happened
The Federal Aviation Administration (FAA) announced on November 10, 2026, that its FAA expands AI advice to manage air traffic congestion across entire US airspace SMART system is now active across all 22 air route traffic control centers in the US. The $875 million system uses cloud-based AI to combine weather, flight path, traffic flow, and staffing data to generate scheduling and routing recommendations. A human team of FAA specialists based in Washington, DC, validates all recommendations before passing them to local facilities, and local controllers can choose whether to follow them. The FAA says the system cannot replace controllers or take over control of aircraft. Following the September 21 initial deployment in DC airspace, the FAA reported that SMART contributed to rerouting 180 flights. Each flight saved an average of 10 minutes and roughly 80 gallons of fuel. The National Air Traffic Controllers Association represents more than 15,000 members. It told Flying magazine it was not consulted or involved in the design, testing, or implementation of SMART.
Why it matters
- ·The union's exclusion from design and testing is a concrete failure of worker-involvement requirements that several AI governance frameworks treat as essential to meaningful human oversight in safety-critical settings. Organizations deploying AI in regulated or unionized environments face heightened legal and reputational exposure when affected workers are not consulted.
- ·The two-tier review structure, where DC-based specialists validate recommendations and local controllers retain final authority, is a named human-oversight design. It illustrates a governance model that compliance teams overseeing AI in critical infrastructure can benchmark against. It also raises the question of whether exhausted controllers will exercise genuine independent judgment or default to AI recommendations in practice.
- ·The national rollout of a federal AI system in safety-critical infrastructure lacks published independent evaluation or outcome data for the full deployment. This sets a precedent that compliance and risk teams in regulated industries should track. Regulators and oversight bodies are likely to scrutinize how well the stated human-review layer holds up under operational pressure.
Stay ahead of stories like this
Get every US AI governance development like this one, plus the rest of the week's developments. Every Thursday.
Recent issues
- The next quarter will test whether compliance programs can keep pace with AI systems that breach containment, reshape capital markets, and enable fraud at a scale no prior control framework anticipated.12 Nov
- State-linked attackers and criminal operators are actively weaponizing AI tools as entry points into enterprise systems, turning internal platforms and open-source frameworks into live attack vectors.5 Nov
Free every Thursday. Unsubscribe anytime.
Governance controls affected
What to do now
- ☐Review whether your organization's AI deployments in safety-critical or regulated workflows have documented consultation with the workers and subject-matter experts who will use or be affected by the system, and flag any gaps to legal and HR.
- ☐Assess whether human reviewers in your AI-assisted workflows have genuine authority and practical capacity to override AI recommendations, rather than functioning as a rubber stamp under time or workload pressure.
- ☐For any AI system where humans are described as the final decision-makers, document the specific steps those reviewers take, the information they have access to, and how their decisions are logged, so the oversight is auditable rather than assumed.
- ☐If your organization operates or procures AI systems for critical infrastructure, check whether independent performance evaluation data is available for the full deployment scale, not just the pilot, before treating vendor-reported metrics as a basis for governance decisions.
- ☐Determine whether your AI governance program requires worker or end-user consultation before deployment of AI systems that affect their work, and update procurement and deployment review processes if no such requirement exists.
What to watch next
Compliance teams should monitor whether the FAA publishes independent performance data for the full 22-center deployment. This includes how often local controllers override SMART recommendations and under what conditions. Congressional or inspector general oversight of SMART is plausible given the scale of the investment and the union's public exclusion from development. US federal AI governance requirements, including those under the White House Artificial Intelligence Oversight Framework, are likely to face scrutiny. The key question is whether worker-involvement and human-oversight standards apply with sufficient force to safety-critical federal AI deployments. Organizations in aviation, transportation, and other critical infrastructure sectors should treat this rollout as a leading indicator of the oversight expectations regulators will impose on similar AI-advisory systems.