Who is developing in-time safety—and what has actually been demonstrated? Explore named programmes, their AI role and the next milestone worth watching.
Sentinel’s opening assessment. In this sample, NASA and the FAA-sponsored ResilienX work provide direct IASMS evidence. EASA develops assurance guidance; CAAS has a named AI trial target; Boeing reports AI use in safety analysis. These are different contributions—not a global league table. Open each record for sources and limitations.
10 of 10 programme records · Sources checked 26 September 2026 · Not a live feed
2027 hurricane-relief tests were planned; completion is unverified
NASA reported a simulation with 12 drone pilots managing up to six beyond-visual-line-of-sight drones, and said related hurricane-relief flight tests were planned for 2027. The System-Wide Safety project page later stated the project had concluded and would no longer be updated.
Professionally reviewedRecord revision 2Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-27 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Predictive risk support and human-autonomy teaming are research themes; the test report does not establish that every tested function uses machine learning.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A substantial direct IASMS research foundation in this source sample. Historical demonstrations must remain distinct from current programme activity.
What this does not prove
Neither these experiments nor the archive notice establish worldwide deployment status. Previously announced future tests need fresh confirmation.
Next evidence to watch
Whether a successor conducted the announced tests, plus published scope, results and operational transition evidence.
Milestone / event date
NASA reported 2024 drone tests and a 5 March 2025 simulation; its System-Wide Safety project was marked concluded on 16 February 2026.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Revision 2 drafted from new or corrected primary-source evidence; held for owner professional review.
2026-09-27 · Owner approved revision 2 for publication after professional review of the evidence summary, classification, limitations and publication suitability.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
United StatesDirect IASMS
FAA-sponsored · ResilienX-led team
Regulator-sponsored research / technology supplier
Security and performance monitoring for UTM
TechnologyTest results reported
Completed contract report; follow-on status unconfirmed
The contractor report describes FRAIHMWORK monitoring service health, integrity and performance, with contingency management and selected failure-mode tests in a UTM ecosystem.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Automated monitoring and mitigation are explicit. The source report does not establish a specific machine-learning role for the whole system.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
More concrete than a concept announcement: a named system, test scope and documented failure scenarios. FAA sponsorship is not blanket FAA operational approval.
What this does not prove
Contractor-authored evidence hosted by FAA; test results do not prove independent safety benefit or nationwide implementation.
Next evidence to watch
A named operational deployment, its approval scope, ongoing performance and independent evaluation.
Milestone / event date
Final report for FAA contract 697DCK-23-C-00287; exact publication date not established from the report.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Primary sources rechecked. No material change to this record; its approved content is retained.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
EuropeEnabling capability
EASA
Aviation safety agency
AI assurance and human–AI teaming guidance
Policy & guidanceProposed guidance
Proposal evidenced; final adoption not established
EASA has proposed an AI trustworthiness framework and further concept-paper guidance. These are evidence of regulatory development, not approval of a particular AI system.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Assurance of aviation AI, including AI assistance and human–AI teaming.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
An important assurance track to watch alongside technical demonstrations. Rules and technology progress answer different questions.
What this does not prove
A consultation or proposed issue is not a final rule. This record does not claim the 2025 consultation is still open.
Next evidence to watch
Final adopted material, applicability dates and documented application to a named aviation use case.
Milestone / event date
NPA 2025-07 announced 10 November 2025; proposed concept-paper Issue 03 dated June 2026.
The programme combines safety reports, flight, traffic and weather data for shared analysis. The reviewed page describes development through 2026 and a planned operational phase.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Big-data and data-science analysis are explicit; a particular AI model or autonomous mitigation capability is not established here.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A relevant foundation for shared safety intelligence. Its institutional and data-governance work matters as much as analytical tools.
What this does not prove
The page includes a January 2024 update. Reaching the target year is not proof that the transition happened.
Next evidence to watch
An explicit operational-phase announcement and delivered analytical outcomes with evaluation details.
Milestone / event date
Development began September 2022. Programme page describes a target transition in 2026.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Primary sources rechecked. No material change to this record; its approved content is retained.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
United KingdomEnabling capability
UK Civil Aviation Authority
Aviation regulator
AI-enabled occurrence-report processing
TechnologyPilot testing reported
AI sorting and labelling of Mandatory Occurrence Reports; outcomes not published
The CAA says it is testing whether AI can sort and label Mandatory Occurrence Reports faster and more accurately. The public update gives no measured accuracy, error rates, human-review controls or safety outcomes.
Professionally reviewedRecord revision 2Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-27 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
AI-assisted processing of occurrence reports; operational scope and performance results are not specified in this source.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A practical SMS intelligence use case worth following. The reported pilot activity is not an evaluation or an approved operational capability.
What this does not prove
The public update reports testing but is not a pilot evaluation or implementation certificate.
Next evidence to watch
Evaluation of sorting quality, error rates, human review and any operational use; distinguish a trial from an approved safety capability.
Milestone / event date
CAA commitment in March 2026; CAA AI page reports testing in Spring 2026.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Revision 2 drafted from new or corrected primary-source evidence; held for owner professional review.
2026-09-27 · Owner approved revision 2 for publication after professional review of the evidence summary, classification, limitations and publication suitability.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
SingaporeEnabling capability
Civil Aviation Authority of Singapore
Regulator / air navigation service provider
AI Multi Sector Planner and Digital ATCO Assistant
TechnologyDevelopment; trials planned
Multi Sector Planner trial target: 2027; read-back error detector live trials planned for later in 2026
CAAS says it has developed an AI Multi Sector Planner for trials from 2027 and is developing a Digital ATCO Assistant. It also says its automated pilot-controller read-back error detector is due to enter live trials later in 2026. No trial start or outcome is established here.
Professionally reviewedRecord revision 2Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-27 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Controller support: planning across sectors and anticipating traffic or weather conditions. These are decision-support initiatives.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A clearly identified, forward-looking AI programme with a stated trial target. It belongs on the watchlist, not in a list of completed IASMS deployments.
What this does not prove
The announcement does not supply trial outcomes or establish a system-wide IASMS implementation.
Next evidence to watch
Confirmation of either trial starting; operating boundaries, controller workload, read-back error rates, fallback behaviour and safety-assurance results.
Milestone / event date
CAAS announcement: 22 July 2026. Multi Sector Planner trials from 2027; read-back error detection live trials later in 2026.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Revision 2 drafted from new or corrected primary-source evidence; held for owner professional review.
2026-09-27 · Owner approved revision 2 for publication after professional review of the evidence summary, classification, limitations and publication suitability.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
EuropeEnabling capability
Airbus
Aircraft OEM / digital services
Skywise aviation data analytics
TechnologyProduct capability described
OEM describes available AI analytics; deployment scope unverified
Airbus describes Skywise as a data platform using AI to analyse information from disparate sources.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
AI-based data analysis. Specific deployed models, safety-critical uses and assured performance are not established by this overview.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A relevant industry capability, but a product description cannot settle how it performs in a particular operator’s SMS.
What this does not prove
OEM self-description, not independent validation or evidence that Skywise is a complete IASMS.
Next evidence to watch
A named operator case study with deployment dates, safety use case, accountable decisions and outcome evidence.
Milestone / event date
Current Airbus product overview accessed 26 September 2026; no dated implementation milestone established.
Boeing reports using advanced analytics and machine learning to support teams identifying potential product hazards, including analysis of in-service data.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
Hazard identification and decision support for safety specialists.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
A concrete example of AI within safety work—not just an autonomy vision. Evidence quality must remain visibly distinct from independent evaluation.
What this does not prove
The manufacturer’s own account does not quantify a causal safety improvement or establish complete IASMS deployment.
Next evidence to watch
Published evaluation methods, error rates, human oversight and independently examined safety outcomes.
Milestone / event date
Chief Aerospace Safety Officer report reviewed 16 September 2026; individual AI deployment date not stated.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Primary sources rechecked. No material change to this record; its approved content is retained.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
AustraliaRelated activity
Civil Aviation Safety Authority
Aviation regulator
RPAS and AAM Strategic Regulatory Roadmap
Policy & guidanceRoadmap published
Late-2024 review captured in January 2025; no later activity status verified
CASA’s roadmap sets out regulatory development, systems assurance and safety/resilience work for remotely piloted aircraft and advanced air mobility. CASA describes its website as the status source, but the checked activity page still shows a 30 January 2025 update.
Professionally reviewedRecord revision 2Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-27 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
No specific operational AI safety capability is established by this roadmap record.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
Useful context for the environment in which future IASMS capabilities may operate. AAM preparation alone is not IASMS adoption.
What this does not prove
The PDF and online activity page both predate this review by over a year. Roadmap targets do not establish delivery, current activity or IASMS adoption.
Next evidence to watch
A fresh update to the living roadmap pages and any named IASMS or AI safety-assurance trial with documented results.
Milestone / event date
PDF version 2.1, January 2025, captures the late-2024 review. The online airspace activity page was last updated 30 January 2025.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Revision 2 drafted from new or corrected primary-source evidence; held for owner professional review.
2026-09-27 · Owner approved revision 2 for publication after professional review of the evidence summary, classification, limitations and publication suitability.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
GlobalEnabling capability
ICAO
International standards and guidance body
Safety Intelligence Manual — Doc 10159
Policy & guidanceGuidance published
Manual described in official guidance overview
ICAO describes guidance on collecting, processing, analysing and using safety data and information.
Professionally reviewedRecord revision 1Source check: 2026-09-26Next check due: 2026-10-05Professional review: 2026-09-16 · Aviation-safety editorial review of the baseline evidence summary, classification, limitations and publication suitability. Not regulatory approval or independent certification.
AI ROLE
A foundation for safety intelligence, not an AI implementation or authorisation.
Sentinel assessment & source record
Sentinel’s assessment · AI interpretation
The organisational intelligence discipline remains relevant whether analysis uses AI or conventional methods.
What this does not prove
A manual does not prove State implementation. National applicability and controlled text must be checked separately.
Next evidence to watch
Documented national implementation and practical, evaluated uses of AI within safety-intelligence governance.
2026-09-16 · Existing Global Watch baseline carried into the tracker; no fresh source verification claimed.
2026-09-16 · Revision 1 professionally reviewed and approved for publication. Programme facts and source-check dates unchanged.
2026-09-26 · Primary sources rechecked. No material change to this record; its approved content is retained.
Sentinel’s interpretation is AI-generated and professionally reviewed before publication.
READ THE STATUS, NOT THE HYPE
One programme. Several separate questions.
We track evidence, not a single maturity percentage. Technical progress and policy progress are different tracks. Programme activity is separate from the highest milestone documented.
The source explicitly identifies IASMS work or tests. This does not, by itself, mean operational approval.
Enabling capability
Relevant AI, analytics, assurance or safety-intelligence capability—not proof of a complete IASMS.
Related activity
Adjacent AAM, UTM or regulatory work whose IASMS implementation link is not established.
Evidence & freshness
Each record distinguishes event, source and review dates. An official plan supports a planned milestone; an OEM claim remains self-reported. “Not verified” is not “not happening.”
NEXT IN THE RESEARCH QUEUE
Coverage is deliberately unfinished.
This first edition is a selected public-source baseline, not a worldwide census or a ranking of safety performance. These organisations are queued for programme-level verification, not labelled as laggards.
MalaysiaCAAM
Not yet assessed in this edition
ChinaCAAC and Chinese OEMs
Not yet assessed in this edition
United Arab EmiratesGCAA and AAM operators
Not yet assessed in this edition
JapanJCAB and industry partners
Not yet assessed in this edition
OEM / supplier expansionCollins Aerospace and Honeywell
Not yet assessed in this edition
No real-time surveillance, accident prediction or automatic website updating is running here. Research findings require review before this dated record changes. Explore the permanent research archive →