AI Counter-Programming

Big tech companies want to convince you that we are at the precipice of AGI, touting their new capabilities to pump up stock price. They come out with new products bragging about their ability to replace and automate human reasoning. Somehow they miss that the rampant fraud, scams, and gaming-of-the-system that they claim cannot be taken down undermines this claim.

Not a week goes by when I don't get multiple connection requests on Facebook (as my first example) from an existing friend. It's the same name, the same profile picture (or one of their photos), many of the same person details (location, etc.). Since I'm already a friend, I know it's a scam, usually sent to their entire friends-list. I report it.

About half the time, Facebook sends me a note letting me know that they've taken down the profile. The other half, they let me know that this person has not violated their terms of service, but thanks for reporting nonsense and please go away. It's SO obvious of a scan that a 6-year-old could identify it, but Facebook, a company claiming to have cutting edge inference capability cannot.

This isn't a case where brilliant criminals are meticulously skirting the edges of a carefully curated perimeter to sneak through tiny holes in the fence. The scammers aren't even really trying. The photos aren't doctored in any way - no noise introduced or alteration. The names aren't ever-so-slightly altered to confuse a number one with a lower-case "L". It's obvious fraud and Facebook could take it down in an afternoon.

To make this obvious, here's a first-pass decision-tree that would stop this activity cold:

1. Does this name already exist?
2. Is the photo associated with an already-created account?
3. Are any of the other details copied from the already-created account?
4. If "yes" to the first 3, then delete the account, block the MAC/IP address

Forget about AI - this is a well understood, easily solved problem. I can completely understand AI being used to sort through much more complex and nuanced issues, but even the most basic checks are ignored.

I remember when Google searches were useful, generating the best result in the first place, even often using the "I'm Feeling Lucky" button on Google's front door (the 2nd button that nobody seems to remember) to take me to their algorithmically-optimized perfect choice. Those days are long gone, with several ads followed by several dubious SEO-optimized results yielding a search within a search.

On Amazon, I can find hundreds of e-books that are not written by the famous authors that, on first glance, one would think they were. Often these are AI generated pulp with similar or identical titles of popular best-sellers, and often significantly discounted from the original publication. They are designed for one clear reason - to dupe voracious readers into buying something if they don't inspect their purchase too closely before closing the sale. Using a remarkably similar decision-tree, Amazon could stop this in seconds.

Then there are the vendors in the Amazon marketplace sending low-quality knock-off products with misleading brand-accurate photos and description, sometimes taking months to arrive from Asia. If they arrive.

I can go on, but you get the picture. Every one of these companies has publicly claimed or testified before congress that they are working on the fraud issues, working on removing hate-speech or child porn, but the problem is difficult and must be manually addressed by carefully curating humans. Yet everything else is fair game for AI.

What Meta, Google, and Amazon don't seem to realize is these fraudulent activities fundamentally damage their claim that AI is ready for prime-time. They don't implement it themselves, so why should I?

The truth, I fear, is more insidious: Fraud that benefits the company is exempt from AI.

When Meta gets new accounts, even bots, they serve more ads and make more money. Every misleading Amazon e-book generates revenue.

It feels like either adult supervision is absent at the most senior level of these companies or they are simply terrible bad stewards of their brands and the promise of great technology.

Interested in taking your business to the next level? I consult and coach on technology, operations, marketing, business development, and overall business strategy.

Talk to me!