How To Spot An AI Ethics Watermelon

The apple might be the forbidden fruit that caused Adam and Eve to crash out of the garden of paradise, but for digital ethicists like myself – we learn to beware of watermelons.

A watermelon has a beautiful green shell, which when cut reveals a yellow crust shrouding a deep red fleshy inner.

So too, can organisations be watermelons.

A watermelon in the ESG space is a company that looks ‘green’ on the outside, but look too deeply below the surface and all sorts of risks become apparent.

25 years ago, it was impossible to tell the difference between companies who had the interests of wider society at heart from those who just had a photo in their brochure of their annual charity run.

By 2021 the asset management industry has matured such that fund managers demand data from ESG Market Data providers who provide ratings and scores as to the ‘green’ credentials of an organisation. Commitment to carbon neutral by 2030? Great! – but prove how you are going to operationally deliver that. With ESG ratings, you can smell a watermelon from afar.

While I was at Fidelity, I collaborated with a long-term friend in the US on an idea of how to apply ESG methodologies to the question of AI ethics. Now, just over a year after starting EthicsGrade – we’re busy rating companies not on their environmental credentials as such, but rather how mature their response to the questions posed by ‘the techlash’ are.

So how to spot an AI ethics watermelon?

Here’s my top tips:

  • A company that has published their AI ethics ‘principles’ but has scant resources online on how the people affected by data driven decisions can have confidence in the protocols that ensure those principles are followed-through.
  • A company that has appointed a “Head of AI ethics” or other such role without the sufficient resources to be able to make for change.
  • A company where the buzz on social/ blogs/ news sites towards their AI ethics efforts feels desperately at odds with their core values statement.
  • A company that sponsors a lot of events on AI ethics, but really just wants to talk about its clever ML Ops platform (which, is really not anything to do with ethics).
  • A company that isn’t willing to recongise that it might be poisoning its ecosystem while at the same time as purifying its own activities.

We’re now just 3 weeks away from the next milestone in the long process towards the regulation of AI driven by the European Commission. I’ll be writing more over the coming weeks about what to expect in these new regulations and what organisations need to do to prepare themselves (and what we, as consumers, need to demand from the brands we love and trust).

For now, just look out for those watermelons. When you start looking, you’ll see them everywhere!

Original post:

Leave a Reply

Your email address will not be published. Required fields are marked *