Is This the Age of Irresponsible AI — or Just Irresponsible Claims About It?

Is the age of irresponsible AI upon us? Or just the age of making irresponsible claims about it?

This week, it was reported, that Delta Airlines has started leaning on AI to offer dynamic and personalized pricing to customers (The article from Fortune is linked is the comments). The strong implication of this is that in our AI adoption rush corporations like Delta are now jumping into throwing your personal information at AI models with abandon just to make a few extra bucks at checkout.

Nightmare scenarios could ensue. What if the AI determines that Black families would be more willing to pay a higher ticket price than white families to take their annual vacations? (We know that certain AI implementations have had race-based biases before). Or what if the AI believes that supporting elderly or disabled passengers costs the airline more and so they should pay a premium rate? (We know that considerations of equity and fairness are high on the list of issues where AI could have detrimental impacts). How could we possibly ensure these systems are transparent? And how do we know what personal information they are absorbing and from where?

The reality is though that in reviewing the service offerings from the partner that seems to be driving this technology for airlines, Fetcherr, the opportunities currently seem far less sophisticated or scary. They have a comprehensive blog post (linked in the comments) about their services to the airline industry, and it appears they have tools that adjust pricing contextually but not personally on an individual level. And while there is nothing to say this couldn't be a longer-term roadmap, in any regulated industry there would be an uphill struggle trying to implement technology that works in this way.

The most likely explanation for what is going on here is that Delta has bought into an AI provider to more rapidly and efficiently do what it has already been doing on pricing, but like so many corporations, wants to maximize its "AI-ification" to its shareholders, and so has pushed the messaging as far as it could reasonably go.

This in itself though is a form of AI-based irresponsibility. Whether connected with the products being promised to customers, or the savings being sold to shareholders, we are in the midst of an AI over-promise explosion. And no matter how good AI's performance is, there is a widening gulf between what it can actually achieve in the short term and the promises that have been made.

So, is Delta about to start squeezing more money from your wallet based on discriminating attributes about you personally? Probably no more than it already has, but it already knows a great deal about your preferences based on purely contextual factors. Truly static airline pricing is long gone, and the ability for any vendor to change up pricing based on a multitude of contextual signals is becoming much easier in the AI age.

Posted on Linkedin on 07/20/2025 -> View Linkedin post here