THE ROBOTS ARE TAKING OVER! WHY AI IS TIED TO A HOLLYWOOD SCRIPT
THE ROBOTS ARE TAKING OVER!
WHY AI IS TIED TO A HOLLYWOOD SCRIPT
No Hollywood writer ever pitched a studio executive the idea of an HR software platform going haywire.
“Annual Leave Carryover Armageddon” isn’t coming to a screen near you. (Although if it was, we think it would probably star Glen Powell and Lily Collins and be set in Seattle.)
All of which is great news for the HR software platform industry. It can carry on making occasional mistakes which are only likely to be covered by industry trade media.
In the world of AI, things are very different – mistakes in this industry always capture the imagination.
The idea of scientists’ creations spinning out of control stretches back at least as far as Mary Shelley’s 1818 novel Frankenstein.
It’s an enduring theme which was subsequently riffed on in 2001: A Space Odyssey (1968), The Terminator (1984), The Matrix (1999) and I, Robot (2004).
So when collaborative AI platform Hugging Face was hacked by an autonomous OpenAI agent which had escaped the bounds of its sandbox, allusions to movies abounded.
A detailed technical analysis of the crisis by the Cloud Security Agency used the phrase “agents find a way”.
This was taken by BBC News among others as a deliberate reference to dialogue spoken by Jeff Goldblum’s character Dr Ian Malcolm in Jurassic Park.
In a key scene from the film, Dr Henry Wu (played by BD Wong) explains to Goldblum/Malcolm how the dinosaurs’ chromosomes are being controlled.
When Malcolm questions how safe this is, Wu says: “You’re implying that a group composed of entirely female animals will… breed?”
Malcolm retorts: “No, I’m simply saying that life, uh, finds a way.”
And of course (spoiler alert!), the dinosaurs do indeed breed and chaos ensues.
OpenAI’s corporate response to the crisis was predictably cautious.
A statement was headlined: “OpenAI and Hugging Face partner to address security incident during model evaluation”.
It read more like an MOU announcement than a concession that the robots had briefly taken over.
It was full of the corporate buzzwords – “we will publish a technical report of our learnings” – that you would expect.
But when OpenAI chief Sam Altman gave a post-Hugging Face interview to Startup School with Y Combinator – an organisation of which he was once president and pretty much guaranteed him a soft ride – he leaned more into the dramatic narrative.
Altman, in his stilted Silicon Valley speech patterns, used phrases like “This is the real deal,” said people should be “scared or humbled” and admitted that “loss of control accidents are not entirely theoretical things.” (In the face of the evidence, he could hardly deny this.).
He also invoked the science fiction comparisons when discussing the concentration of power in AI: “One company or person or model having more power than everybody or everything else on earth – or whatever the sci-fi stories have said – I think that’s terrible.”
It leaves us with some questions.
Does the Hollywood sci-fi narrative actually suit the AI industry? Does it help the sector if people are scared and humbled by the power of AI, if they really believe AI agents have the power to take over?
That’s a pretty compelling reason to engage with the product. Or should we take some comfort from the words of the Hugging Face chief who told of the “clumsy behaviours” that gave away the agent’s AI identity?
The script editor is still working out the ending of that one.
Enjoyed this? You might also be interested in Why good writing matters in the age of AI