Out there at Technology Review I found this article about an “AI” (you may raise your eyebrows here) which is supposed to better in stock-market speculations than actual humans. For a brief introduction let me quote TR:
It’s called the Arizona Financial Text system, or AZFinText, and it works by ingesting large quantities of financial news stories (in initial tests, from Yahoo Finance) along with minute-by-minute stock price data, and then using the former to figure out how to predict the latter. Then it buys, or shorts, every stock it believes will move more than 1% of its current price in the next 20 minutes – and it never holds a stock for longer.
TR points out, that analyses similar to the described algorithm exists since the 90ies. However, the new systems doesn’t actually parse all the data, but concentrates on some keywords which seem to be of relevance.
I see two very odd flaws there.
- A good AI predicting the stock-market based on human-written text – which could be technically used by *anyone* who could afford it – would lead to a situation where stocks keep heating up. Speculation will grow rapidly and positive feedback loops will possibly run into an overdrive situation. I wouldn’t opt in for a ban on such a software but on full disclosure if this software was used on a certain bid. This could help in debugging situations and to give legislators something to think off when the shit already hit the fan.
- If the algorithm actually concentrates on keywords in context rather than in the whole analysis of the text, I bet a fiver that it wouldn’t even take a few weeks until some clever consulting company analyzed the algorithm and makes up a process how to tweak your fiscal reports so that AZFinText favours this text. Think of the stock-market equivalent of a Google bomb.
Nobody in Technology Review’s forum seems to be worried about the real-life implications… I think I’m just pointing out the obvious and that the stock-market professionals already made up their own ideas.