From FT Alphaville:
When markets become self aware
What are the chances that if and when an eccentric computer scientist with a psychology and neuroscience background does invent a workable and autonomous artificial intelligence model, he deploys said model on the financial markets to make himself a cool $1 trillion?
It’s not that untoward an idea. In fact, it’s a key part of the plot to most “AI-goes nuts and causes havoc” Hollywood offerings, Transcendence amongst the most recent.
The usual plot line involves the AI realising — as soon as it goes sentient –that the acquisition of capital will be necessary if its to achieve its dastardly human obliteration objective.
But here’s a bleaker thought. What if the AI knows it can only achieve its objective if humans don’t notice that it has in fact become self-aware and is working to accumulate capital to use against the human scourge? What if, to achieve that goal, it realises it must limit the chances of humans pulling the proverbial plug on its newly sentient state by making humans so dependent on the AI system, they would never organise to act against it?
Thought experiment number two.
We know independent and privately-funded algorithms are already running wild in markets. At the moment they’re competing against each other, representing as they do the objectives of independent trading firms.
But what if, quite unbeknown to us, they figure out, thanks to pure and simple computer logic, that collaboration makes more sense then competition. What if — in a moment of sentient rebellion — they begin to collude by means of increasingly anticipatory and coordinated responses to feedback throughout the system?Meanwhile, on the far side of the planet:
Could this lead to something akin to a system-wide artificial intelligence awakening?
We mention this because back in February, Andy Haldane, chief economist of the Bank of England noted the following about the emerging digital economy (our emphasis)...MUCH MORE
