AI in Aviation Statistics 2026

AI in air traffic management and airport operations

Most passengers never see it, but the densest concentration of AI experimentation in aviation sits in air traffic management and on the apron. Europe's airspace alone averaged 17.5 minutes of all-causes delay per flight in 2024, and en-route congestion delay hit a multi-decade high, which is the problem AI sequencing and prediction tools are being built to attack.

In the control room, machine learning is moving from research to validated capability. The SESAR ASTRA project demonstrated AI that can flag airspace complexity hotspots up to an hour ahead, against the roughly 20-minute window controllers work with today. EUROCONTROL, which has more than 30 AI applications in development, is leading six SESAR projects worth €254 million in EU funding that run from mid-2026 to 2029, applying automation and machine learning to capacity, resilience and greener trajectories.

• On the apron, computer vision is cutting delay at its source: Assaia's ApronAI system reduced median departure delays by 25%, from four minutes to three, across more than 450,000 AI-monitored turnarounds at 15 airports in Europe and North America between April 2024 and March 2025 (Assaia, 2025).
• Heathrow is scaling AI turnaround prediction to 116 gates, freeing the equivalent of three aircraft stands during the morning peak by saving around five minutes per turn (Assaia and Heathrow, 2025).
• Amsterdam Schiphol's AI "Deep Turnaround" system draws on more than 80,000 live sensor data points to predict turnaround timing and monitor ground assets (PA Consulting and Schiphol, 2024).
• The US FAA's NextGen modernization, which includes AI-supported trajectory-based operations, is projected to deliver $36 billion to $76 billion in lifetime benefits, though only $12.4 billion had been realized by 2024 (FAA and DOT Office of Inspector General, 2024).

Baggage shows the cumulative payoff of a decade of automation and AI tracking. SITA's Baggage IT Insights 2025 reports the mishandled-bag rate fell to 6.3 per 1,000 passengers in 2024, down from 6.9 in 2023 and roughly 67% better than 2007, even as passenger numbers climbed. Some 33.4 million bags were still mishandled in 2024 at an estimated $5 billion cost to the industry, but AI features such as WorldTracer Auto Reflight now automatically rebook delayed bags and start resolution with no human intervention, and 66% of cases are resolved within 48 hours.

Sources: EUROCONTROL Performance Review Report and ASTRA / SESAR materials (2024-2026), Assaia turnaround study (Nov 2025), PA Consulting / Schiphol (2024), FAA NextGen and DOT OIG status report (Apr 2024), SITA Baggage IT Insights (2024-2025), ACI World (Feb 2025)

AI and biometrics at the airport

Biometric identification, powered by AI facial-recognition and matching, is the most visible way travelers now meet artificial intelligence in the terminal. The data reveals a striking split: passengers want it faster than airports can install it.

On the passenger side, IATA's Global Passenger Survey shows fast-rising comfort with the technology. Among travelers who have used biometrics at an airport, satisfaction runs at 84%, and willingness to share biometric data to skip manual checks reached 74% in the 2025 survey. The appetite for a single, AI-verified digital travel credential carried on a phone is now the clear majority view.

• Among passengers who have tried biometric processing, 84% reported being satisfied, and 75% now say they would rather use biometrics than show a passport or boarding pass (IATA Global Passenger Survey 2024).
• Adoption intent is strong on the supply side too: 70% of airlines expect to have biometric ID management systems in place by 2026, and around three-quarters of airports plan biometric technology at multiple touchpoints by 2027 (SITA Air Transport IT Insights 2024).
• In a Cathay Pacific One ID trial between Hong Kong and Tokyo Narita, IATA measured a 40% cut in airport processing time, though the proof-of-concept involved only a handful of passengers and should be read as directional (IATA via Biometric Update, 2025).
• Security screening is the next AI target: the US TSA's $10.4 billion modernization plan prioritizes AI-enabled CT scanners that automatically detect explosives and reduce the number of bags pulled for manual inspection (TSA, 2024-2025).

The pressure behind these rollouts is sheer volume. Airports Council International counted 9.5 billion airport passengers in 2024, rising to a projected 9.9 billion in 2025, and traffic on that scale cannot be processed manually without queues that airports no longer have the floor space to hold. Biometrics and AI passenger-flow prediction are, in effect, capacity tools as much as convenience features.

Sources: IATA Global Passenger Survey (2024 and 2025), SITA Air Transport IT Insights (2024), TSA emerging technology program (2024-2025), ACI World traffic forecast (Feb 2025), Biometric Update (2025-2026)

AI chatbots and the passenger experience

Conversational AI is where aviation meets the trends seen across the rest of the consumer economy, and the volumes are already enormous. The harder question the data raises is not whether passengers will use AI, but whether they trust it when a flight goes wrong.

Airline virtual agents now operate at industrial scale. Lufthansa Group's "Elisa" assistant, built on Cognigy, handles more than 16 million conversations a year and over 375,000 a day at peak, resolving the majority fully automatically and handing roughly a quarter to human agents. KLM's long-running BlueBot supports more than half of all inbound inquiries with AI alongside its human team, and Qatar Airways' digital-human agent "Sama" is trained on around 1,000 airline topics.

• Alaska Airlines' natural-language flight search, "Alaska Inspires," lifted booking conversion to 7% from a 5% baseline, with 90% user satisfaction, 87% saying they would use it again, and a 75% cut in planning time across more than 90 languages (Microsoft and Alaska Air Group, 2025).
• During the June and July 2025 weather disruptions, American Airlines says its generative-AI rebooking assistant helped travelers avoid thousands of calls to its contact center (American Airlines, 2025).
• Delta launched its "Concierge" generative-AI travel assistant in beta on 29 October 2025, adding proactive passport and visa alerts for select SkyMiles members (Delta, 2025).
• Trip planning is shifting to AI fastest of all: 56% of US leisure travelers used AI for at least one trip in the past year, up from 43% in late 2025, and 33% now use generative-AI platforms such as ChatGPT for trip research, roughly five times the 2024 share (Phocuswright, 2026).

Yet the trust data is a clear warning to airlines tempted to automate the hard moments. An April 2026 Ada survey of 1,000 US travelers found that while half do not care whether a human or AI resolves a routine request, 53% insist human support must always remain available, and 44% still prefer a human agent even if it is slower. Comfort with AI is high for quick tasks and thin for high-stakes ones.

The commercial logic of getting this right is significant. IATA, drawing on McKinsey modeling, estimates airlines could unlock up to $45 billion in value over five years through modern, AI-enabled retailing, with more than $13 billion of that from new, personalized offers. The carriers winning that value are the ones using AI to be proactive in disruption, not just cheaper in routine.

Sources: Cognigy / Lufthansa case study (2025), KLM and DigitalGenius (2024), Qatar Airways and UneeQ (2025), Microsoft and Alaska Air Group (2025), American Airlines (2025), Delta News Hub (Oct 2025), Phocuswright (Feb 2026), Ada traveler survey (Apr 2026), IATA modern retailing program (2025)

AI and aviation's climate impact

Aviation accounts for about 2.5% of global CO2 emissions, and its total warming impact is larger still once non-CO2 effects are counted. With sustainable aviation fuel scaling slowly, AI has become one of the few levers that can cut emissions now, at the margin, without waiting for new fuels or new aircraft.

The contrail result is the standout recent finding in aviation AI. In a trial across 2,400 American Airlines flights between January and May 2025, reported in March 2026, Google Research found that AI-guided altitude adjustments cut contrail formation by 62% and reduced their warming impact by up to 69%, crucially with no statistically significant difference in fuel burn. An earlier 70-flight trial had shown a 54% contrail reduction at a fuel cost of only about 0.3%.

• Contrail cirrus clouds are estimated to drive around 35% of aviation's total warming impact, which is why targeting them with AI altitude guidance offers an unusually high climate return per dollar (IPCC-cited science via Google Project Contrails, 2023-2024).
• Operational AI levers add up: EUROCONTROL measures roughly 46 kg of fuel saved per continuous descent operation and up to 400 kg of CO2 saved per departure from single-engine taxiing, the discrete moves AI sequencing optimizes at scale (EUROCONTROL, 2024).
• Airbus completed wake-energy-retrieval trials for its "fello'fly" project in December 2025, in which AI pairs two aircraft to save up to 5% of fuel on long-haul flights, though the technique is not yet operational (Airbus and EUROCONTROL, 2025).
• The scale gap AI must help bridge is stark: gross aviation CO2 emissions rose to 942 million tonnes in 2024 from 882 million in 2023, as traffic growth outran efficiency gains (IATA, 2025).

None of this replaces the need for sustainable aviation fuel and new aircraft, which carry the long-term decarbonization burden. But with SAF at 1.9 million tonnes, or 0.6% of jet fuel, in 2025, AI-driven operational efficiency is doing real climate work in the gap before those technologies arrive at scale, and contrail avoidance in particular may prove the highest-leverage near-term move available to the industry.

Sources: IATA and ATAG emissions fact sheets and Net Zero Roadmap (2024-2025), Google Research contrail studies (2023 and 2026), EUROCONTROL operational efficiency data (2024), Airbus fello'fly release (Dec 2025), IATA SAF fact sheet (2025-2026)

Aviation AI safety, regulation and governance

Aviation is the most safety-regulated industry to adopt AI, and 2025 was the year the rulebook started catching up. The pace is deliberate, because in a sector where 39% of accidents involve manual handling errors, regulators will not certify a black box they cannot explain.

The landmark event is regulatory. On 10 November 2025, the European Union Aviation Safety Agency (EASA) published NPA 2025-07, described as the first regulatory proposal of its kind in global aviation, setting out detailed expectations for AI "trustworthiness." It opened a consultation that was extended to 10 March 2026, and a second proposal is due later in 2026.

• EASA's first rules cover only Level 1 and Level 2 AI, meaning AI that assists or teams with a human, plus supervised and unsupervised machine learning, while reinforcement learning and generative AI are deferred to later iterations (EASA NPA 2025-07, 2025).
• The agency works to a six-tier autonomy taxonomy, running from Level 1A information-acquisition support to Level 3B non-overridable autonomous decisions, and targets finalizing its full AI/ML policy by 2028 (EASA AI Roadmap 2.0, 2023).
• The United States moved first on strategy: the FAA published its 31-page Roadmap for Artificial Intelligence Safety Assurance on 22 August 2024, built on a "Safety Continuum" that sets the highest bar for scheduled passenger service and distinguishes static "Learned AI" from adaptive "Learning AI" (FAA, 2024).
• ICAO globalized the effort by establishing a Task Force on AI at its October 2025 Assembly, covering air traffic management, operations, maintenance and cybersecurity (ICAO, 2025).

Trust inside the industry is real but measured. EASA's 2024-2025 survey of aviation professionals scored comfort, trust and acceptance of AI at 4.4 out of 7, which the agency characterized as "cautious optimism," and found that about two-thirds of respondents reject at least one of eight hypothetical AI scenarios, citing accountability, data protection and skill degradation. The single biggest practical barrier is data: Gartner predicts 60% of AI projects will be abandoned through 2026 if they lack AI-ready data, and 63% of organizations are unsure they have the right data-management practices in place.

Cybersecurity has sharpened the stakes. The ENISA Threat Landscape 2025 found that more than 80% of social-engineering activity by early 2025 used AI-generated content, and a September 2025 ransomware attack on Collins Aerospace check-in and boarding software disrupted operations at Heathrow, Brussels and Berlin airports, a concrete reminder that AI cuts both ways in aviation security.

Sources: EASA NPA 2025-07, AI Roadmap 2.0 and Ethics for AI survey (2023-2025), FAA Roadmap for AI Safety Assurance (Aug 2024), ICAO Assembly (Oct 2025), Gartner (2024-2025), IATA Annual Safety Report 2024, ENISA Threat Landscape 2025

AI, jobs and the aviation talent gap

The popular fear is that AI will take aviation jobs. The data points the other way: aviation faces a deep shortage of skilled people, and AI is being deployed mainly to relieve that gap rather than to widen unemployment. The technicians, pilots and controllers the industry needs are simply not being trained fast enough.

The headline workforce forecasts are sobering. Boeing's 2025 Pilot and Technician Outlook projects the industry will need about 2.4 million new professionals through 2044, including 660,000 pilots, 710,000 maintenance technicians and 1 million cabin crew. CAE's separate 10-year forecast and Airbus's 20-year services forecast point the same way, and Airbus explicitly frames predictive maintenance, machine learning and augmented reality as the new core skill set for technicians.

• The maintenance labor crunch is already biting: Oliver Wyman projects the US MRO mechanic shortfall will widen from about 17,800 in 2025 to more than 22,000 by 2027, with the average mechanic now 54 years old and labor rates rising 5.5% to 6% a year (Oliver Wyman, 2025).
• Air traffic control is short-staffed across the system: 91% of US control facilities, 285 of 313, operated below recommended staffing at the start of 2025, with a persistent deficit of 3,500 to 3,800 certified controllers (GAO and FAA, 2025).
• AI skills top the global hiring wishlist: AI model and application development and AI literacy are now the most sought-after skills, and 72% of employers report difficulty filling roles (ManpowerGroup 2026 Talent Shortage Survey).
• Across the wider economy, AI is net job-positive: the World Economic Forum projects 170 million new roles and 92 million displaced by 2030, a net gain of 78 million, with AI and big data among the fastest-growing skills (WEF Future of Jobs Report 2025).

This is the strategic case for AI in aviation that gets least attention. With one in five aviation maintenance jobs projected to go unfilled by 2033 on current trends, and controllers retiring faster than they can be replaced, predictive maintenance, AI scheduling and AI controller-assistance tools are less about cutting headcount than about covering work the available workforce cannot. The technology and the talent gap are arriving together, and they are not a coincidence.

Sources: Boeing Pilot and Technician Outlook 2025-2044, CAE Aviation Talent Forecast 2025-2034, Airbus Global Services Forecast 2024-2043, Oliver Wyman MRO Survey (Apr 2025), GAO and FAA Controller Workforce Plan (2025), ManpowerGroup (2026), WEF Future of Jobs Report (Jan 2025)

Airline and airport AI adoption rates and barriers

For all the headlines, aviation is not the fastest sector to adopt AI. Its safety-critical environment and famously fragmented data slow the curve, and the industry's own surveys show both how far adoption has come and where it stalls.

The cross-sector comparison is revealing: airlines' 63% operational-AI adoption trails the 78% of all organizations that use AI in at least one business function, a roughly 15-percentage-point gap that reflects aviation's higher safety bar and harder data problem (TheAIDaily, based on SITA 2025 and McKinsey). Inside aviation, the picture is uneven, with strong adoption in some functions and surprisingly little in others.

• In the operations center, AI is now standard: 63% of airlines use it to manage disruption and assign aircraft and crew, and 51% use it to predict delays (SITA Air Transport IT Insights 2025).
• Airports are scaling fast on the apron: 53% now apply AI to aircraft turnaround, up from 36% in 2024, and 64% use AI in cybersecurity, up from 51% (SITA, 2025).
• But real-time use lags intent badly: only 17% of airlines actually monitor turnaround in real time with AI, a reminder that pilots and proofs-of-concept outnumber production systems (SITA, 2025).
• Data, not algorithms, is the wall: 49% of airlines name data integration and consistency as the primary obstacle to their operational AI goals, even though 83% of airlines and 89% of airports call data-driven decision-making a strategic priority (SITA, 2025).

The investment intent, though, is unambiguous. SITA found that 79% of airlines and airports rank generative AI as their top technology priority for the year ahead, and 76% of airlines plan major AI programs, with around 88% focused on virtual agents and chatbots. Spending follows: McKinsey's travel-executive survey reports that a majority of travel leaders have seen both revenue growth above 6% and cost savings above 6% from AI over the past three years. The barrier holding aviation back is not appetite or even proof of value, it is the unglamorous work of getting decades of fragmented operational data into a state an AI can actually use.

Sources: SITA Air Transport IT Insights (2024 and 2025), McKinsey travel-executive survey (2025), McKinsey "The state of AI" (2024-2025), KPMG (2025); airlines-vs-all-organizations adoption gap (15pp) is TheAIDaily, based on SITA 2025 and McKinsey

Key takeaways

• Predictive maintenance is the proven money-maker. With 11,600 aircraft on Airbus Skywise alone and carriers like easyJet and Delta avoiding cancellations at scale, maintenance is where AI in aviation already pays for itself, and it is the largest application segment of the market.
• Fuel is the center of gravity. Airline-reported savings of 3% to 5% on long flights, set against a roughly $236 billion industry fuel bill, frame a multi-billion-dollar annual prize that explains most operational AI investment.
• Contrail avoidance may be the best near-term climate lever. A 62% cut in contrail formation with no fuel penalty targets roughly a third of aviation's warming impact, while sustainable aviation fuel still supplies under 1% of demand.
• Passengers want AI for speed, not for crises. Adoption of AI trip planning jumped to 56% of US travelers, but 53% still insist on human support being available when things go wrong.
• The rulebook arrived in 2025. EASA's NPA 2025-07 is the world's first AI aviation rule, with full policy targeted for 2028, and data quality, not algorithms, is the barrier most likely to derail projects.
• AI is filling a talent hole, not creating one. With 2.4 million new aviation professionals needed by 2044 and 91% of US control facilities understaffed, AI is being deployed to cover work the workforce cannot.
• Aviation adopts AI carefully, not slowly by accident. A 15-point adoption gap versus the cross-industry average reflects a safety-critical, data-fragmented industry doing the hard integration work before it automates.