The dangers of fake AI

Visit the High Level Logic (HLL) Website
— Home of XPL (eXtensible Process Language)
It's like XML, but you can store source code in it.
What's worse than media hype about robots taking over the world? Ideas that over-hype the character and promise of AI have slowed its progress several times. Irresponsible journalism leads to great expectations that cannot be fulfilled fast enough. People lose faith in the effort. Research funding dries up. Entrepreneurs can't find interest. Now we have a new villain; con artists selling fake AI.

All the big companies, like IBM and Apple want to be seen as leaders in the field. The marketing departments of even those companies with humble technological beginnings like Facebook, a “social media company”, Google the search engine company and Microsoft, the perpetual follower want in on the deal. It suggests that they're on the cutting edge. Even people like Elon Musk want to be players. But wait, there's more.

If you've applied for a job in the modern era, you might have done so by filling out a long form on the Internet. It's then possible that you didn't have a chance because you never got to talk with a real human that had a clue what the job was actually about – one that you were perfectly suited for. That's an example of automating beyond the abilities of modern software (and programmers, and the people who planned and paid for the system).

The fact that a lot of automated software doesn't work well hasn't stopped the developing companies from selling it or customer companies from using it – no matter how much it hurts them. I think the above example is a good one. An automated system that delivers a biased and narrow view of which applicants should be interviewed; based merely on the fact that the software in the process is dumb, costs the hiring company a lot. They don't end up with the right people to do the work, grow, and prosper – certainly not to be competitive.

So what solution can the marketing department in Crap Software, Ltd. find to fix the problem of having crappy software that isn't really up to the job? The popular answer nowadays seems to be by pretending that it's AI. Not that the people in the marketing department know what that is really. But it's a popular buzz word that suggests something new – not that same old crap they were trying to sell last year. Except often enough, that's what it is.

From automating personnel software to working out who should be kicked off of Facebook to various schemes in customer processing and satisfaction ideas and onto a range of other human decisions that require effort; there are liars out there trying to sell “solutions” by pretending that the same old disappointing software they were selling last year is now smarter than you. Or at least that it's now suddenly, magically, up to doing a competent job so that humans don't need to – even though it's not.