Jeppesen ForeFlight says it’s harnessing AI to bring plain language NOTAM summaries to its products. The ClearNOTAMs feature will be available following AirVenture 2026 and will “translate dense, heavily-coded NOTAMs into clear, plain-language summaries, helping pilots grasp critical flight information faster and with greater confidence,” the company said in a pre-AirVenture release. All existing individual ForeFlight customers will be able to download a free trial of the feature after the show. There has been widespread criticism of the growing number and complexity of NOTAMs, and the FAA says it has heard the complaints and will be addressing it through an ongoing NOTAM modernization program.
The company says ClearNOTAMs doesn’t just decode the acronyms and technical abbreviations. It also allows pilots to sift through and prioritize them based on the specific trip they’re planning. “Every pilot across the globe has experienced the challenge of sorting through pages of NOTAMs before a flight,” said Cole Crawford, Director of Product at Jeppesen ForeFlight. “ClearNOTAMs helps pilots spend less time deciphering information and more time understanding it. By transforming complex NOTAM language into clear, concise summaries, we’re making one of the most important parts of flight planning more accessible and actionable.”
The company says it uses AI models specifically trained to accurately unravel NOTAM jargon to get at the heart of what they’re trying to convey. After feeding the NOTAM in, the system issues a clear English summary of the NOTAM, but pilots can also view it side by side with the original to double-check its accuracy. The software makes those summaries available throughout the various ForeFlight functions so when it’s handy to refer to the NOTAM it’s available on that page. The company said pilots who used beta versions of feature during testing reported feeling more confident about getting the NOTAMs prior to flying.
AI has taken a toll in all aspects of daily life.
I work in the automotive finance and leasing world and get to speak with a lot of customers who show high levels of frustration with AI controlled phone robots. By the time they finally get to speak with a human – anger and frustration are at peak levels.
These days even the dumbest redneck is capable of penning long and well written legal letters (obviously assisted by Chat GPT) and we in turn have software that allows us to summarize and condense the wall of text into key points – upon which, with no legal case pending yet, a oneliner response is written.
Various publishing outlets have removed all human touch from writing and content, Youtube, Instagram, Facebook all chuck full of fake stuff that occupies the flatheads for hours each day.
Most of this stuff is fake, some of it is dangerous and those able to profit from it milk the cow while its still able to stand upright. Not much will be left by the time we realize that automation and artificial intelligence has never had any other purpose than to replace humans and individualism.
I am 51 and probably one of the younger ones still actually writing and researching for my content. Already a dinosaur by this day and age and for some stuff I have to use 99.9% of my brain to filter AI generated stuff out or verifying the validity of the information.
Thats how we wanted it to be and here we have it. I will not live to see the end of this idiocy nor will our generation benefit from any of it. Maybe the next – or the one after that.
Don’t know – don’t care.
Racist insults like “redneck” and “flathead” (not sure what that means…AI says it’s an insult dating back to the 18th century — who knew?) are used by most people to signal that they are intelligent, whereas the hoi polloi is not. Statistical principles say that only half of the population can be of above average intelligence, and therefore that about 4 billion earth-dwellers must be below average.
I find that no matter who you ask, whether educated or not, they will attest to their own intelligence — and above average intelligence to boot. My empirically-sourced study reveals that 100% of people consider themselves to be of above average intelligence!
Anyway…this educated pilot is sitting around right now waiting for the AC repairman to come so I can resume living in a cool house. I hope he’s at least of average intelligence, redneck or not.
“Statistical principles say that only half of the population can be of above average intelligence”
Statistical principles do NOT say this. Given the set “1,1,1,1,10”, the average is 3 (15/5). The term you are looking for is ‘median”, not “average”.
Sorry for the rant.
Thanks for the correction. Would human intelligence (definition..?) follow bell-curve distribution? If so, wouldn’t my point about average intelligence be valid?
See Dunning-Kruger (ask wikipedia, or heaven forbid, AI).
I think the American philosopher Forrest Gump offers the best reference: “stupid is as stupid does”.
Yes, I think so, but I don’t know if anything like that has been measured. The results would be interesting to me.
With a nod to RichR’s comment, a list of sub-median intelligence pilots could start with those planning to transit NYC airspace tomorrow.
That’s in fact correct. IQ value is not an absolute metric. It’s a relative value, where 100 is set to be a median value and the rest are distributed across a perfectly symmetrical bell curve, making the median exactly the same as average.
Construing a racist insult into the words redneck of flathead (own word creation describing individuals easily entertained by fake content on mobile devices) is pretty strong. But okay for me too. Have a great day. AC repairmen are generally of above average (median) intelligence.
Garmin Pilot is ahead of Foreflight in that it already sorts NOTAMs by type (aerodrome, runway, FDC, etc..). On a day VFR flight I can ignore FDC notices and zero in on categories relevant to my flight. I assume Garmin will implement the AI summarizing feature and look forward to such a development.
Well said Mr. Baker.
I would say decoding NOTAMs is one GOOD use of AI. Every industry and occupation has its special “language” to keep outsiders thinking they’re too ignorant to know what the “expert” knows. This has no place in aviation, however, due to the large number of recreational pilots flying around. Anything that enhances safety is a good thing.
Not many of us like AI in general, but this could be pretty useful. Our NOTAM system is a relic of the old teletype days, and it’s ridiculous that while being awash with weather data, pilots have to wade through them and hope they didn’t miss something. Having AI help interpret NOTAMs is a good idea; having the FAA issue them in plain English so they don’t need interpreting would be even better.
I recall similar concerns expressed by folks in the early 2000s as Google search was powering up. The derisive term used was “Google University”, suggesting that any novice could qualify themselves as an expert by using Google for a particular topic. AI is in a similar state today when used to understand generalized topics. For something more focused, it likely will be more accurate. For decoding NOTAMS, I expect high fidelity because there is a limited, constrained set of terms that need to be translated. Sorting out which NOTAMS are relevant to one’s route of flight, I expect more inaccuracy, so would use it, but then try to verify. Verifying that the included NOTAMS are relevant is relatively quick and straightforward. Verifying that no relevant NOTAMS were left out, however, that’s much more time-consuming. It may end up providing no time advantage. Nonetheless, I’ll try it and see. If it ends up being too messy, I’ll step back and let it evolve.
For anyone that hasn’t experienced flt w/TFRs, and thinks non-compliance is just down to carelessness…feel free to plan a flt for 7/4/2026 transiting NYC area.
I know there is about to be a major change/update in the NOTAM system. Any idea how much the new system will actually clean up the original NOTAMs before they even get to ForeFlight?
The changes in ForeFlight will be a welcome change. But it would also be nice if the Gov system was cleaned up too. It is fine to keep the Teletype coding for those areas around the world that need it. But there’s no need for the codes here in the US any more.
Like Mr.Jason J Baker,
I work in the automotive world including finance. When any new AI driven application launches, its initial roll out is programmed on a certain amount of accuracy but is designed to gather and analyze all the data that it can find. Similar to self learning EFI added to classic cars. However, one must drive the car…a lot…under a wide range of conditions to learn all the nuances of that particular engine under a variety of circumstances…weather, temps, engine timing, cam specs, etc to enable it’s data gathering facilitating its ability to improve its performance. Most initial comments about performance is negative because the owner only drives it on nice days, under the most ideal circumstances, preventing the system from “learning”. So, baseline tune allows the car to start and run. But optimum performance comes from lots of drive time under a variety of conditions, gathering, storing, then applying the data for improved performance. It would be cost prohibited or next to impossible to program a self learning system with everything it needs to know for optimum performance. It launches with enough info to start the data mining. It can only improve with use.
People get upset about auto insurance payout quotes or a finance turn down from a lending institution on a loan application…saying something like “they are trying to screw me.” AI is handling your loan application. AI is handling your damage claim. Today, ” they” is a computer with an AI program. Don’t take it personal. You are arguing with a microprocessor because you don’t fit within the predetermined and present goalposts for an approval. As it gathers more data, including every keystroke you have already put into this global financial system, maybe you will in the future. No one is looking over your credit report. Only AI is making that decision based on not only your credit report, but with virtually everything you have purchased, every tax return, every banking activity, online job function that involves the use of a computer, app, cell phone, social media, etc. AI is self learning about you. AI is also learning how to be predictive about you, including spending habits. Part of your loan approval is based not only on your present bureau score, but sorting out the likelihood you will pay your debt.
So, AI will be helpful in summarizing a NOTAM. But it will have mistakes that must be “relearned” for correction because it might not have all the accurate data to come to the right conclusions initially. And the incorrect data leading to the wrong conclusion does not go away. It gets sort of squeezed out among all the new data that it is continuously processing. Hopefully among this avalanche of new data, the correct data is also there. Sometime it is, sometimes it is not, sometimes things are intentionally left out because it might just be not good for the corporation’s bottom line, or in the NOTAM example, the FAA. So, truth becomes potentially arbitrary. Predictive AI will eventually determine its version of truth. Is it really true? Inquiring minds would like to know.
This is exactly where AI can shine. NOTAMS are based on fairly rigid rules. Flight planning is also based on rules. It is virtually certain that, right out of the gate, AI will be more accurate in translating NOTAMS into colloquial language, and more importantly, sifting through the ones that aren’t applicable for a specific flight and showing those that are, in order of significance (or chronologically, as they pertain to flight path). This is where computers in general are superior to humans, and modern AI shines.
That said, there’s always a chance of error or omission, and then, product liability might rear its very ugly head.
Notams are decades past needing to be converted to a user friendly format, both the digesting of them and the appearance. And the ways to do so have been available for years. My AC repairman could do it I’m sure. As a pilot for all the years and as a controller, I never have been able to easily digest what was presented. And, because of that, ashamed to say, basically avoided them. For a local flight, just didn’t care. For our local ATC serviced airport, our staff would pull them out as necessary and type them friendly into a preduty read and initial binder. For a cross country, too damn many about antennas and lights, so just ignored them. I was wrong to do so, but they’re wrong to make them so unfriendly. So there.
It might take a few tries……