There’s an interesting Freedom of Information (FOI) response by the Information Commissioner’s Office (ICO) on the website WhatDoTheyKnow. In response to the question
have you examined the use of AI to help you in doing your work as an organisation?
their reply includes the statement that
For information, the ICO does not use any artificial intelligence (“AI”) technology.
However, if one uses most of the standard definitions of AI (such as the one from the government’s National AI Strategy: “machines that perform tasks normally requiring human intelligence, especially when the machines learn from data how to do those tasks”) one might find that hard to believe. What about spam filters on the ICO email network? Or the fact they recommend Google Maps for anyone needing directions to their offices? Or their corporate use of social media? All of those technologies use, or constitute, AI.
There is a wider point here: the task of regulating AI, or even of comprehending how it uses personal data, will fall increasingly on some key regulators in coming years (including the ICO). It is going to be crucial that there is understanding within those organisations of these issues, and if they don’t comprehend now how, within their own walls, the technology operates, they will be starting off on the back foot.
(One should also add that, if the ICO has missed some of its own more obvious uses of AI, then it has probably also failed to respond to the FOI request in accordance with the law.)
The views in this post (and indeed most posts on this blog) are my personal ones, and do not represent the views of any organisation I am involved with.