Quick: what is the FIRST thing that should be open to democratic control, and used for the common good?
Two articles about a great issue of our time just made me a bit sad.
. One is a great piece in which Joi Ito explains how and why “we need social advocates, lawyers, artists, philosophers, and other citizens to engage in designing extended [artificial] intelligence from the outset”. I completely agree with Ito when he says that doing what he proposes may be “the only way to reduce the social costs and increase the benefits of Artificial Intelligence as it becomes embedded in our culture.”
The other article, from the Guardian, is titled “Cambridge Analytica scandal highlights need for AI regulation”, and is about a “Lords report stressing need for artificial intelligence to be used for the common good”.
Source: Google search results for Artificial Intelligence dystopia
</em></u>
What makes me sad is to see, after (at least) a decade of worldwide OpenGov/OpenData advocacy and projects to “create an unprecedented level of openness in Government” in the USA and elsewhere, assertions that those who should “open their designs to democratic control” are NOT governments, but Artificial Intelligence Engineers.
Yes, I myself have pointed out that these days the ones acting as real governments are, instead, private, for-profit, algorithm-first platforms. I also noticed, two years ago, that Brexit also is a wake-up call for the Open Data Movement. Still, titles like the Medium one, or that whole Guardian article, feel a bit too much like surrendering to the idea that, these days, it is Matrix or Skynet who are ruling and, above all, must continue to rule.
Random comments about “Artificial Intelligence for the Common Good”:
- the problem is not what AI does, is what those in power can do with AI (cit.)
- please count up to some billion of billions before saying “smart contracts” or “blockchain”. There is too much hipe around those buzzwords already
- for some reason, the very idea of designing “Artificial Intelligence with controls for the common good” feels a bit too similar to “security through obscurity” in software designing
- maybe the way to go to make AI working for the common good is to have as little AI as possible, as NON-centralized as possible
- Ito’s call to “engage social advocates, lawyers, artists, philosophers etc…” should be applied to Open Government and Open Data projects, even before than to AI ones