Is AI About to Die from Exposure?

Apple announced at its World Wide Developer Conference several new features that claim to be AI. Most of these features are small bore and run on Appel’s own model, but apple did announce a partnership with OpenAI. It is a very unusual partnership, with Apple not paying OpenAI a thing for the use of its tools. OpenAI, is in fact, being paid with every writer’s worst nightmare: exposure.

Apple is not paying for the use of ChatGPT in its products. It is merely going to distribute the features and OpenAI will apparently sign-up lots of people to its services based on that exposure. That seems extremely unlikely, especially since Apple’s own functions will be available for the price of the device. Payment in exposure, as almost every writer knows by now, is essentially no payment at all. And this is not the first time that OpenAI’s deals have appeared much more significant than they were in reality. Most of the money announced in the Microsoft deal was in the form of credits for cloud services. Now, that is not nothing, but it is also not cash or stock. You have to wonder if the next AI winter is coming.

Imitative AI is expensive. It not only requires massive data, which entails storage costs, but massive amounts of computing power. It requires a ton of energy and water to make happen. No Imitative AI company is profitable, at least as far as we can tell, and OpenAI’s CEO is admitting that the next version of ChatGPT is not a significant advancement over the current models. Salespeople for Google and Amazon have dialed back expectations for the usage of these tools, perhaps because their customers aren’t finding them that useful. Investors are noticing and scaling back their exposure to these firms. It very much does look like snow in the AI world.

And that is probably a good thing. Imitative AI is based on the theft of people’s work in an attempt to profit by replacing them. It produces unrealizable results, spreads disinformation, and makes people worse at their jobs. I am sure that they will make some money, but their business proposition appears to be based on the idea that imitative AI is inevitable and therefore you should use it now, no matter how much it gets wrong. It is, from a business perspective, largely a scam, in other words.

I look forward to the coming AI winter, honestly. I have spoken about this before, but imitative AI offends me as a technologist. I cannot find the person who first said this (because AI has helped make Google completely unreliable) but it sums up the problem with modern tech extremely well: I don’t want AI that writers and draws for me, leaving me to do the dishes and laundry. I want AI that does the dishes and laundry for me so that I have more time to write and draw. The fact that the people in charge of these firms are doing the opposite pretty much tells you all you need to know about their concern for the rest of society.

Because there are many, many problems that technology can help solve, and many actual kinds of drudgery that technology could eliminate for us. But none of them offer the venture capitalists a return to the golden days of easy, explosive growth. They tried with crypto, and with NFTs, and now with imitative AIs. They went for the quick buck rather than the buck made by being helpful to society.

If winter is coming to imitative AI, I hope it snows long and deep. Maybe, just maybe, we can get technology that helps people solve problems more meaningful than how can I not pay people to create.