The stories about AI chatbots are getting ridiculous. Worse, they’re getting unavoidable in the headlines. Could be the last dance of the pretenses of the human brain, a cosmic gift for idiots – AI does everything? No, it doesn’t. It does some basic stuff well, complex stuff often badly, and it runs on sequential logical rails where people’s minds don’t.
The historical image of thinking machines is old, and the image is hardwired into media. Thinking machines are somehow godlike. They’re infallible. They can do anything people can do. The naivete is almost total. So, more than a bit annoyingly, is the assumption of naivete in the market.
That’s the image; this is the reality:
- Erratic, semi-tested tech which may or may not work.
- Ridiculous, obviously scripted anecdotes of “sentience” and even more hackneyed cliché encounters on the subject.
- Logical non-sequiturs of emotional behavior in AI to the point of tedium. So what? So nothing, so far.
- AI deepfakes which are basically criminal.
- See any major business assets in this mess you want to spend millions on?
A great ripoff in progress? Probably.
The reality of AI is much less interesting and far more turgid. Most of what you see in the noisier headlines is basically promotional material. It’s great, it’s colossal, it’s fabulous, it’s fantastic… No, it isn’t. It’s just highly publicized.
Not much of what Bing’s Sydney and Bard are doing is new.
- It can write essays? AI has been doing that for years. What I’ve seen from OpenAI is definitely competent, but this isn’t so much new as AI writing V 2.0.
- It’s some sort of oracle? Why? AI achievements tend to be about running data, not conversation.
- Chatbots are gods? Hardly. They’re more like a phone queue in practice. The difference between “Press 1” and conversational sequences is almost nothing.
Try thinking of it like a smart phone. What is a smart phone? It’s a phone. It has other tech tacked on. Most of the actual tech on a phone is decades old, descendants of older tech. AI in this form could be described as “responsive software” at best. Not new, not all that different, just more conversational.
Scientific AI usage is almost the exact opposite of these bots. It can run tests, do the logic, etc. and you rarely hear a word about it. A chatbot, however, is some sort of major breakthrough? Since when?
…Which is where the probability of ripoffs emerges. Like smart phones, AI will be heavily overpriced. Its functions will be more Pokémon Go than Data from TNG. The trouble here is that nobody’s setting performance parameters. “It’s a great car; one day it will have wheels, a chassis, and maybe even an engine. Seats optional.”
AI in its current much-hyped form couldn’t be much less impressive. What has it actually done? Nothing useful has really been delivered. Is the world, or are the claustrophobic minds of slavish tech-addicts any better for these months of drivel? Not noticeably.
So set some standards. Physical and performance parameters, not pabulum, are what are required. Accuracy and positive benefits will say a lot more than this amateur hour direct marketing ever will.
Let’s see something more credible than paying big money for things that don’t actually do much and talk far too much. Real AI will come, but not like this.
____________________________________________________________
Disclaimer
The opinions expressed in this Op-Ed are those of the author. They do not purport to reflect the opinions or views of the Digital Journal or its members.
