🏳️‍⚧️ trans rights are human rights 🏳️‍⚧️
Theme

The FTC Is Suing Air AI, But Can It Stop the Next AI High-Tech Scam?

They Called It the Future. It Was a $19 Million Trap.

Source Filed: August 25, 2025 • FTC v. Air AI Technologies, Inc. et al.
FTC Action AI Fraud Business Opportunity Scam Telemarketing Violations

While their customers were taking out high-interest loans and maxing out credit cards to buy into the AI gold rush, the founders of Air AI were secretly broke, had given up their apartments, and were desperately trying to raise outside capital just to stay afloat.

The Setup: How They Sold a Dream to People Who Could Least Afford to Lose It

Air AI Technologies, Inc. and its web of interconnected shell companies did not build a real business. According to the FTC complaint filed August 25, 2025, they built a sales funnel. The company targeted entrepreneurs and small business owners, the exact people most vulnerable to a pitch promising financial independence through technology, and used social media ads, relentless phone calls, texts, and emails to pull them in.

The product had two main forms. The first was the “Air AI Access Card,” a bundle of coaching materials and software sold for between $15,000 and $30,000 (between one-third and two-thirds of the median American household’s yearly income). The second was a “Licensing Business Opportunity,” which gave buyers the right to resell the same AI technology to other businesses, priced anywhere from $25,000 to $100,000 (up to twice what a median American family earns in a year).

Defendants told every single prospect they were a “perfect” candidate for the program. The FTC’s complaint makes clear this was a script, applied uniformly regardless of the consumer’s actual situation, background, or financial position.

The AI Was Broken Before They Even Sold It

The centerpiece of the pitch was an AI agent named “Odin,” which Defendants claimed could conduct 10 to 40 minute human-sounding phone conversations, with “infinite memory, perfect recall,” and the ability to act across 5,000-plus applications without any training, management, or motivation. That is a remarkable claim. The FTC says it was also a lie.

In documented testing, Odin read portions of scripts aloud that were never meant to be spoken. It hung up on callers seconds after they answered. It struggled to write down a basic email address. When asked questions it didn’t have scripted answers for, it simply made up responses on the spot. One license-holding consumer spent months trying to make the system work for a building company, attempting to configure the AI to answer questions about floor plans and schedule follow-up meetings, before concluding he could not possibly show the product to a real client.

A separate software tool called “Sherlock,” promised as part of the package, was never built at all. It did not exist. Consumers paid for it anyway.

“At worst, Defendants sell consumers a junk suite of services that do not exist or are not consistently available.”
— FTC Complaint, paragraph 7
$0 $25k $50k $75k $100k $15k $30k $25k $100k Access Card (Min) Access Card (Max) Licensing (Min) Licensing (Max) Price Charged to Consumers Air AI Access Card Licensing Business Opportunity
What Air AI charged consumers. Individual prices were set based on the consumer’s apparent ability to pay, not any tiered feature structure. Source: FTC Complaint, paragraphs 34, 91–96.

A Shell Game Built Across Six Companies

Air AI Technologies was not operating alone. The FTC complaint names six corporate entities, all sharing the same Phoenix, Arizona address, which appears to be a co-working space, commingling over $23,631,000 (the equivalent of the median American household saving their entire income for over 470 years) across shared bank accounts. Defendants Caleb Matthew Maddix and Ryan Paul O’Donnell controlled all six entities simultaneously.

The structure was designed to make accountability difficult. Employees listed as contractors for one entity received payments from another’s bank account and had access to yet a third company’s funds. This is the corporate architecture of obfuscation, built to make it harder for anyone, regulators included, to follow the money.

The Non-Financial Ledger: What Money Can’t Measure

There is a story the dollar figures do not fully tell. The people who bought into Air AI were not reckless. They were hopeful. Many were small business owners, the exact demographic that mainstream financial institutions have historically underserved, looking for a legitimate edge in a world where AI was being sold as the great equalizer. They saw the ads. They heard the sales calls. They read contracts that promised, in writing, that they would get their money back if the product failed. They made a rational decision based on information that was, according to the federal government, systematically false.

One consumer was told he could not afford the $30,000 price tag. Defendants responded by helping him take out a loan for $15,000 and structuring the second $15,000 as a cut of his future sales, business that never materialized. He finished the transaction with $15,000 in debt, no launched business, and no refund. The FTC’s complaint notes that Defendants “never expected him to come up with the rest of the money,” implying they already knew the product would fail him and treated the reduced price as a way to extract whatever they could. That is predatory by design.

A second consumer told Defendants he couldn’t afford the Access Card at all. Defendants convinced him to put down a $2,500 deposit, the only amount he could scrape together, and then take on a $17,000 loan to cover the rest. The refund guarantee was the tool they used to make him comfortable with the debt. He heard “risk free” and believed it. That trust was the transaction Defendants were actually completing.

A third consumer, a woman, already had a $15,000 personal loan out to pay for the program. Defendants then called her repeatedly to pressure her into opening a new credit card with a $4,500 limit and spending it on additional services. When she resisted and refused to charge the full amount, they negotiated her down to $1,000 on the new card. The FTC’s complaint documents this as a pattern: Defendants “appear to set the price depending on a consumer’s ability to pay, extracting as much money from each individual consumer as possible.” That is not a business model. That is predation measured in credit limits.

And then there was the waiting. Consumers who asked for refunds, many of whom met every contractual criterion, were strung along for months. Defendants promised funds were on the way. They said refunds would come as soon as possible. They cut off access to the software those consumers had already paid for, leaving them unable to even try to make the system work while simultaneously refusing to give back the money. Some consumers watched Defendants continue to sell new licenses to new people during the same period they were being told there was no money left to honor their refund.

One consumer put it plainly in the FTC complaint: he grew “concerned that they would just be raising money from other people like me to pay out my refund.” He had watched a business opportunity transform, in real time, into something that looked structurally identical to a Ponzi scheme. That realization is its own kind of damage. It is the damage of learning that the system you trusted was built to take from you, and that the people running it knew it all along.

“Consumers enter contracts with Defendants believing that their investment is essentially risk free. Some consumers even take on Defendant-brokered loans from third-party lenders in order to cover the purchase price.”
— FTC Complaint, paragraph 9

Legal Receipts: Their Own Words, in Evidence

The following are direct quotes and factual statements drawn verbatim from the FTC’s complaint, filed August 25, 2025. These are the words Defendants used in contracts, emails, and marketing materials, now entered into the federal record.

“Air AI is backed by the industry’s boldest guarantee. Within 6 months of joining the program (the ‘Starting Term’), the Client will have used Air AI to generate revenue equal to double or more of the initial cost of the Air Access Card. If not, Client will have the option to cancel their Air Access Card and forgo the Air platform upon which they will be refunded the full amount of their initial cost. Furthermore, Client may receive an unconditional refund within 72 hours of purchase for any reason.” — Direct contractual language from Air AI Access Card agreements, as quoted in FTC Complaint, paragraph 80
“Our company has come into some intense financial hardship. I mentioned before that we were looking to raise a large amount of outside capital and that it would hopefully put us in a better position. Sadly, our intense and persistent efforts to raise outside capital were not fruitful. This has led to us being in an even worse financial position than before… We as owners have put 100% of our personal capital into keeping things afloat. We have spent 100% of our cash, maxed out our credit, and even given up our apartment to cut back… At this point… we are not even close to profitable and we have spent every last dollar that we have as owners. We are completely out of cash at this moment.” — Email from Defendant Ryan Paul O’Donnell to multiple consumers, August 2024, as quoted in FTC Complaint, paragraph 86
“We’ve organically had companies reach out to us that have over a hundred thousand reps. So if you had a rep like that, that’d be like ten million dollars a month paid out to you. But let’s see. If you brought something that was twelve thousand reps you’d be making a million dollars every single month forever based off that.” — Defendants’ marketing video claim, as documented in FTC Complaint, paragraph 73(I)
“Despite claiming to be entirely without funds, Defendants continued to sell their Licensing Business Opportunity to consumers.” — FTC Complaint, paragraph 87
“Defendants do not provide any written or audio disclosures or disclaimers limiting the representations made about earnings potential.” — FTC Complaint, paragraph 76
Trustpilot Reviews: Air AI 0 20 40 60 78 3 1-Star Reviews (Negative) Positive Reviews Number of Reviews
Trustpilot review breakdown for Air AI. 78 one-star reviews versus 3 positive, as documented in FTC Complaint, paragraph 100. The public had already sounded the alarm.

Societal Impact: Who Actually Pays When AI Grifters Win

Economic Inequality: The Scam Went Hunting Where Desperation Lives

Air AI explicitly targeted entrepreneurs and small business owners, the segment of the population most likely to be operating on thin margins, most hungry for a competitive edge, and least protected by institutional safety nets. The FTC complaint documents Defendants actively probing for how much each individual could afford before setting a price, a practice it describes as “extracting as much money from each individual consumer as possible.” This is financial targeting, not just fraud.

The loan infrastructure Defendants built made it worse. Their Midas platform did not just accept credit cards. It brokered high-interest consumer loans at rates up to 24.99%, funneling customers who couldn’t afford the product into debt instruments that would compound their losses long after the software stopped working. A consumer who borrowed $17,000 at 24.99% interest and made no income from the product would owe substantially more than the original price within a year. The company built its own debt trap and installed it inside the sales funnel.

The complaint documents at least four separate private lawsuits filed against Defendants in 2024 by harmed consumers. In at least three of those cases, Defendants defaulted, meaning they simply did not show up to defend themselves, while simultaneously continuing to sell new licenses to new customers. The people left holding bad debt had no functional legal recourse, while the pipeline of new victims kept flowing.

The cumulative commingling of over $23,631,000 (the equivalent of every resident of a mid-sized American town each losing nearly $1,000) across six corporate entities means that even in a successful enforcement action, tracing who has what money will be a significant legal challenge. The shell company structure is, in this context, a wealth-preservation strategy built on the anticipation of legal failure.

Public Health: The Psychological Cost of Financial Betrayal

The source material does not include clinical data on mental health outcomes, and this publication will not overstate what the evidence supports. What the complaint does document is the specific mechanism of psychological harm: Defendants deliberately used the refund guarantee as a tool to neutralize consumer anxiety at the moment of purchase, telling people the investment was “risk free” and, in at least one documented case, talking a customer out of canceling within the contractual 72-hour window by reminding him of that same guarantee. The trust was weaponized at the exact moment someone was trying to protect themselves.

Consumers who then lost tens of thousands of dollars, found themselves in debt, lost access to the software they’d paid for, and received no refund had been systematically deceived at every decision point. The financial loss was preceded by the experience of being lied to, strung along, promised resolution, and then ghosted. That sequence, the hope deliberately maintained and then cut off, is a documented pattern of harm that extends well beyond the dollar amount lost.

The “Cost of a Life” Metric

$23.6M Commingled Across Six Corporate Shells Air AI bank account transferred ~$14M to each corporate defendant. Source: FTC Complaint, paragraph 30. Air AI ~$14M out Apex Holdings Apex Scaling Apex 4 Kids New Life Capital Onyx Capital $23,631,000 Total Commingled
The six-entity shell structure Air AI used to move money. All companies shared the same Phoenix address. Maddix and O’Donnell controlled all six. Source: FTC Complaint, paragraphs 26–30.

What Now: Names, Watchlists, and What You Can Do

The Named Defendants

The FTC complaint names three individual defendants personally, meaning they face personal financial liability alongside the corporate entities:

There is a press release about this story here if you want to learn more by visiting the FTC’s website: https://www.ftc.gov/news-events/news/press-releases/2025/08/ftc-sues-stop-air-ai-using-deceptive-claims-about-business-growth-earnings-potential-refund

I also found this Reddit post about Air AI. There were lots of positive comments which makes me suspect that they’re all fake comments planted there by Air AI. But one of them caught my eye since bro was straight up begging for a way to skirt anti-robo call laws. How can these people not realize that they’re the villains here? https://www.reddit.com/r/SaaS/comments/191yw6v/anyone_worked_with_airai/

Explore by category

01

Antitrust

Monopolies and anti-competition tactics used to crush rivals.

View Cases →
02

Product Safety Violations

When companies sell dangerous goods, consumers pay the price.

View Cases →
03

Environmental Violations

Pollution, ecological collapse, and unchecked greed.

View Cases →
04

Labor Exploitation

Wage theft, worker abuse, and unsafe conditions.

View Cases →
05

Data Breaches & Privacy

Misuse and mishandling of personal information.

View Cases →
06

Financial Fraud & Corruption

Lies, scams, and executive impunity that distort markets.

View Cases →
07

Intellectual Property

IP theft that punishes originality and rewards copying.

View Cases →
08

Misleading Marketing

False claims that waste money and bury critical safety info.

View Cases →
Aleeia
Aleeia

I'm Aleeia, the creator of this website.

I have 6+ years of experience as an independent researcher covering corporate misconduct, sourced from legal documents, regulatory filings, and professional legal databases.

My background includes a Supply Chain Management degree from Michigan State University's Eli Broad College of Business, and years working inside the industries I now cover.

Every post on this site was either written or personally reviewed and edited by me before publication.

Learn more about my research standards and editorial process by visiting my About page

Articles: 1920