Continuing with the technology theme of today’s blogs – but let me begin with two disclaimers. First, as regular readers already know, I am not a fan of AI. Second, this blog involves a family member so there may be some bias here – yet there is a cautionary tale for anyone looking on-line for or trying to sell green/ethical consumer goods – and broader concerns about AI in sustainability. With that, here we go…
My oldest son taught me something surprising about the intersection of AI, social media, retail and sustainability/ethical sourcing this week that shows the power AI currently has over consumers and its inability to distinguish ethically-sourced products from other goods. He runs a small company, selling “ethical, handmade, personally hand-picked shark tooth jewelry” made exclusively from 2-23 million year old fossils (not teeth harvested from living sharks) he digs up or dives for in Florida himself and makes every necklace, earring, etc with his own hands. He is passionate about ocean health and saving living sharks, donating a portion of each sale to marine research and conservation efforts.
AI bots at Instagram – one of his main sources of sales – deemed his products to contain parts of endangered animal species and therefore banned them. His appeal was denied (by those bots), citing the same reason. Hyper-technically, animals that have been extinct long enough to be fossilized could be considered endangered, I guess, but… Clearly, “the ‘Gram’s” AI that controls what consumers see on-line can’t distinguish critical differences between what should reasonably be banned and what shouldn’t. This made me wonder if other products/service are inappropriately banned because they include words or ideas outside of assumed negative contexts. For instance, could social media ban posts for neighborhood summertime lemonade stands or lawn care jobs because an AI bot thinks those are “child labor”?
I write a lot about the power of developing products where sustainability is a key customer buying criterion. But if AI bots prevent ethical products from being seen by on-line consumers because of algorithm or training data failures, those products fail for “artificial” reasons. This makes me question how well AI is trained in sustainability matters – despite assurances from AI solution providers. Companies looking to implement or rely on AI for sustainability, ethical or responsible sourcing uses need to understand the training data set used and conduct pre-launch testing applying nuances relevant to the product/situation.
Our members can learn more about AI in ESG here.
If you aren’t already subscribed to our complimentary ESG blog, sign up here for daily updates delivered right to you.
