Opinion Quelle surprise - the actor who played Neo in The Matrix is wary of the burgeoning developments in machine learning and artificial intelligence.

Speaking to tech glossy WIRED to promote his latest cinema outing, John Wick: Chapter 4, Keanu Reeves had some choice words for The Machines.

During the conversation, Reeves, whose character in The Matrix ends up leading a resistance movement against AI-controlled robots that dominate humanity, was asked about a clause in his contract stipulating that his performances couldn't be digitally manipulated without his permission.

He said this was because a film "in the early 2000s, or it might have been the '90s" added a tear to his face. "I was just like, 'Huh?!' It was like, I don't even have to be here."

The discussion then veered into the topic of deepfakes, where machine learning algorithms can be applied to people's faces and voices to recreate them digitally. While the technology is shockingly powerful, its applications can be vile in the case of deepfake pornography or even dangerous if used convincingly to put words in political figures' mouths.

Reeves commented:

The interviewer suggests "The Matrix just looks more and more wildly prophetic by the day," to which Reeves responds:

The "coporatocracy" arms race to "control" the winning AI already seems to be unfolding between Microsoft and Google - as crap as their efforts may seem now - while Meta is making a landgrab for the metaverse, of which Reeves said:

The writer, Angela Watercutter, then asked if she should be worried about AIs coming for her job.

Interestingly, Watercutter notes that during the article's photoshoot she sees a tweet from a writer saying "one of his clients no longer wants to pay him for his work - because an AI will do it for free (the client will pay a cheaper rate for him to clean up the AI's copy, if he wants)."

She tells Reeves about it and jokes that he's probably right. "He gives me a thoughtful look, and then he gets explicit: Corporations don't care about paying artists. Well, what he actually says is this: 'They don't give a fuck.'"

This is just sad. How dumb does some publisher have to be to think an AI can "write" anything? All it means is that you don't care what the words say as long as there are words there.

While mainstream media has its knickers in a twist over ChatGPT et al, the fact remains that AI can't replace any creative work worth preserving, as author and education blogger John Warner argues in this essay, which is well worth a read.

CNET had to pull the plug on its proprietary AI article generator after a series of howlers. Tesla has been accused of faking demonstrations of autonomous driving in 2016. Nick Cave called an attempt by ChatGPT to write lyrics in the style of Nick Cave a "grotesque mockery of what it is to be human."

This is because an AI does not "know" it is driving along a road just as it does not "know" it is writing a Nick Cave song. It has no causal understanding of anything. Ultimately, AIs of the now are just impressive statistical models that can come close - but not close enough - to outputting something resembling human behavior. Even so, it cannot adapt when the situation changes or it is confronted with information from outside of its training data, which leads to egregious errors.

Such mistakes wiped $120 billion off Google parent Alphabet - and this is a good thing that should be celebrated. Microsoft's AI-powered Bing isn't immune either because, in some ways, AI is all smoke and mirrors. Like Neo in The Matrix, it should be resisted at all costs. ®

https://www.theregister.com//2023/02/15/keanu_reeves_ai/