The last decade has seen the world of finance getting dominated by institutions that make heavy use of high-end technology to gain a competitive advantage on the trading floor. The terms High Frequency and Algorithmic trading are almost interchangeable these days when exchanges allow speed of transactions that go below the millisecond barrier. Clearly this is not a world for human traders anymore.
Since these machines are making multi-million dollar decisions within a split of a second people would like to believe that they actually know what they are doing. It is not surprising that there is a substantial investment from financial institutions in Artificial Intelligence alongside the constant strive to improve performance and robustness of these money making machines.
As a result there are those that see the financial industry as the most prominent sector from which we might see the emergence of a highly intelligent machine. Dr. Alexander Wissner-Gross calls this a “planetary scale intelligence” in his speech at last year’s singularity summit:
Global markets certainly provide a very competitive environment for algorithms to “evolve” and the financial incentive in this particular industry is extremely high.
So is this where a strong AI might emerge?
Personally, I am a bit skeptic about that. I absolutely agree with the argument that there has to be some strong financial incentive behind the development of any Artificial Intelligence. But how much financial institutions contribute to that end?
Finance certainly is adopting occasionally some techniques of machine learning and pattern matching but there haven’t been any significant advancements in AI coming from financial institutions. And if there had been they are kept secret for obvious reasons. Apart from the infrastructure investments in bandwidth, finance is not really contributing to the field of AI in any way other than this: It makes it clear that machines can outperform humans in a field where success and failure are measured with real money. Even people that would otherwise consider AI as science fiction and the playground for computer scientists, can not argue that there might be some value in there after all.