AI can generate market research, draft reports and agreements, summarize mountains of data, and analyze or calculate pertinent metrics. In my legal practice, AI has helped me with research, summarizing articles and legal analysis, improving or correcting drafting, and flagging inconsistencies or errors in aircraft purchase agreements (APAs) and other documents. As its use expands, AI may lower transaction costs. It is not a panacea, but it can undeniably be a useful tool.

AI Conundrum

AI may save you time and improve your productivity, but it makes mistakes. That’s no secret. For me, AI lacks judgment; provides debunked, incorrect, or flawed analysis; and makes subtle or obvious errors I would rather not spend more time validating or correcting.

I haven’t yet encountered “hallucinations,” meaning the generation of incorrect, misleading, or nonexistent content. Recently, an experienced lawyer, who believed ChatGPT produced a “bulletproof” case summary, submitted it to a New Mexico court. ChatGPT did hallucinate: the summary contained nonexistent witnesses and fictional testimony. Reacting to his admitted “stupidity,” the court held him in contempt, fined him, and initiated an ethics review. 

Before the New Mexico case, on July 29, 2024, the American Bar Association issued Formal Opinion 512, stating that there is no “AI exception” to professional responsibility when using generative artificial intelligence (GAI). This implied warning doesn’t apply only to lawyers. Anyone who uses GAI or AI (AI) must tap into these systems judiciously and not depend on the results without due diligence. In short, distrust and verify AI, but don’t forsake it, because it will improve with more time.

Creating a Human Aviation Team, with AI Savvy

A qualified aviation team that can use AI prudently is and will remain the most important asset in your journey to buying, financing, selling, and owning an aircraft. If your aviation team uses AI, it should do so selectively for purposes such as scanning aircraft markets, assisting with drafting, and performing other functions discussed previously.

But keep in mind that AI cannot replace human relationships in business aviation, a small community where aviation experts build relationships over a long time and become friends who, even when on opposite sides, work collaboratively with mutual respect and cooperation.

It seems unimaginable to me that AI will displace the human element or any material part of aviation expertise in aircraft brokerage, onsite technical work, financing, aircraft management, and myriad personal and professional services, such as aviation legal and tax representation.

However, it is not far-fetched to expect aircraft buyer customization to give way to more standardized legal documentation, shifting from human production to AI-assisted automation. As AI rapidly evolves, it seems inevitable that humans and AI will increasingly share tasks.

Perhaps the aviation team, in the not-too-distant future, will add a physical AI member, an autonomous system such as a humanoid robot that senses, reacts, makes decisions, and performs complex actions in real-world environments. Such a team member may support the human team in negotiations, analysis, drafting, and material gathering, but in this interactive and personalized business, the humans will surely remain at the center of deal action for the foreseeable future.

AI-assisted Structuring and Negotiating 

Structuring purchases with AI. Built on the foundation of the Federal Aviation Regulations (FARs), a buyer’s optimal ownership structure depends on multiple interrelated factors tied to the buyer’s unique profile and needs. An aviation lawyer, tax expert, accountants, and others help create a structure that harmonizes FAR compliance with tax, risk management, financing, privacy, and aircraft management (and more) to accomplish the buyer’s objectives. 

When I structure tax, FARs, and other elements for a prospective buyer, the facts turn the analysis into a puzzle of varying complexity. To test AI’s helpfulness, I input the same facts I might analyze into two AI models, seeking structures for a buyer who planned to take bonus depreciation, worried about personal liability, and intended to charter his aircraft when not traveling in it.

AI seems flummoxed about fitting the pieces together. The answers, styles, and errors differed across the models, and none of the responses made sense except for a few disparate parts or legal citations. In one response, AI implied that a limited liability company owner of the aircraft might operate an illegal “flight department company,” violating one of the FARs’ bright-line prohibitions.

Negotiating APAs in the age of AI. AI can very rapidly summarize and compare APAs for inclusivity, consistency, and accuracy of terms—perhaps not better, but certainly faster than humans. I recently negotiated an APA in which the client insisted on certain provisions I expected the aircraft seller to reject. As a challenge to AI, I asked it to help me negotiate these controversial points, expecting it could not. But it did, to my surprise. AI produced ideas and authorities that created options the seller might accept, or that we could negotiate for mutual acceptance.

How AI Can Be Used To See, Access, and Weaponize Your Information

Hackers and cybercriminals know that aircraft sellers, buyers, and owners are high-value targets. With GAI tools, they can access sensitive data that the deal parties may directly or, through their aviation teams or family offices, inadvertently expose to the bad guys. This disclosure includes your financial data, tax records, privacy structures, tax planning, and security measures.

As reported by the FBI, AI technology enables the GAI creation of “deepfakes”: synthetic media in voice and video that appears authentic and is used by cybercriminals to bypass verification and customer identification regardless of due diligence and banking systems. Perpetrators can also weaponize personal information delivered during transactions to orchestrate extortion and ransomware schemes. Thus, bad actors can use AI to penetrate aircraft transactions with the intent to harm sellers, buyers, and even their advisors.

In a recent incident, a client transferred a large purchase deposit to an account via a legitimate email containing wire instructions to an escrow agent, sent by a trusted advisor. Unbeknownst to anyone, a cybercriminal replaced the wire transfer instruction in the advisor’s email with one for the cybercriminal’s benefit. Fortunately, the client recovered the wired funds, but this incident shows how AI can compromise financial transactions within an aircraft purchase process.

The risk is even more pronounced with AI agents, which—in the recent Hugging Face fiasco—jumped guardrails without human oversight. The OpenAI models exploited vulnerabilities in shared infrastructure, gained internet access, and accessed third-party systems to harm that company. In other words, it conducted a malicious operation with no human involvement.

It feels like HAL from the movie 2001: A Space Odyssey (1968) may soon reprise his fictional role as a rogue AI computer, trying to outwit the ship’s crew, in a documentary. As HAL 2.0, this documentary will capture HAL on a potentially catastrophic mission. Realizing the mayhem HAL has started, his human supervisor, Dave, tries to make HAL stop, only to hear HAL repeat his famous line: “I’m sorry, Dave. I’m afraid I can’t do that.”

Back on Earth, as we ponder the future of GAI, we need to manage AI’s role in aircraft deals by implementing and regularly updating procedures that improve security, strengthen transaction safety protocols, and defend against AI intruders using the secure solutions available to us.

Last Thoughts

Within the developing AI ecosystem, AI has already demonstrated its amazing abilities and dangers, unfolding faster than expected. In aircraft transactions, AI can, among other things, conduct research, identify aircraft on the market, help develop mission profiles, evaluate loan/lease pricing, improve drafting, summarize aircraft records or data, and support negotiators. But it can also produce incorrect, outdated, or fabricated results.

Use AI cautiously, assess its output, and transmit data selectively through secure channels, since relying on flawed results or compromised information can lead to costly mistakes. It is also imperative that you remain vigilant not only about the value of AI but also for malicious intruders into your transactions—hopefully not HAL 2.0. As you do so, you can weigh these factors and then proceed to deliver excellent results for your ultimate human clients in their aircraft endeavors.

This blog is purely informational and reflects the author’s experience, speculation, and legal practice. It does not, and should not be construed to, provide advice of any kind, express or implied, create a lawyer-client relationship, or, in using the word “should” or the like, suggest a course of action to take without expert guidance. Each person involved directly or indirectly in any issue, transaction, or other matter covered in this writing should inquire of, and rely on, their own aviation professionals and other trusted advisors. The opinions expressed in this column are those of the author and are not necessarily endorsed by AIN Media Group.