Ethics seems to be the new fashionable thing. There is ethics in AI, which we believe is a lost cause because if bad people want to use it for their own ends, they will. It is also the buzzword of choice among marketing people.
Then there is ethics in big data and privacy, which makes more sense.
The Institute of Chartered Accountants in England Wales (ICAEW) have – for some reason – put together a set of principles – and ethical framework in their words – for financial companies.
Big data worries them because they believe that banks and insurers could use the data they have on a person to favour them or exclude them. If someone has a monitor in their car that tells the company that this is a safe driver, then the company could lower the premiums. Conversely, if a driver has speeding tickets and no monitor, the premiums could go up.
The ICAEW provides some case studies around the use of data, as part of their ethics campaign. In the US, apparently some banks have blocked customers’ accounts and credit cards from buying firearms.
There are many examples of companies using data, in ways that are increasingly Big Brotherly. One company monitors employees’ activities throughout the day and scores them accordingly.
In another, slightly surreal example, an investment manager found that an AI liquidity risk model was significantly outperforming any traditional method and the ethics people at the investment house decided that he couldn’t use it – because no-one could figure out how the AI model did it.
The problem is that the ethics vs competitive edge debate seems to swinging away from common sense.
Should a bank decide if you can buy a gun? The firearms lobby in the US would have a pretty strong opinion about that, you can be sure.
Should a company monitor their employees to see who is performing well? Companies, surely, have been doing that for hundreds of years, just not with such sophisticated techniques.
And should an investment manager not use an AI based model because it would maximise his returns but is deemed unethical on the basis that no-one understands it? Give the model to a ruthless competitor and let him tell you whether or not to use it.
Ethics and an ethical framework around how to use AI is probably mission impossible.
Ethics and an ethical framework around data – however big – is probably a good idea but surely the answer is to let customers choose whether companies can use their data, and how. Which is what various pieces of pointless legislation are designed to do.
The real answer is to implement good old fashioned Common Sense.
Wait a minute, there is probably an algorithm for that.