The male platypus has a venomous spur behind its hind legs. It is one of a handful of venomous mammals on the planet. I think this is not widely known because the platypus is an absurd creature that defies expectation or categorization. The fact that it is also venomous is not the headline. You really need to sit down and have a beer with a platypus before the spurs thing even comes up.
We value rules, structure, and taxonomy so wildly that when something rule-breaking comes along, we focus on categorization and ignore nuance. And sometimes we greet the new thing with skepticism and fear.
When the Internet came on the scene, people did not jump to a nuanced appreciation. It took many years before we dropped the suspicion, and appreciated the value in browsers and email. People cautiously welcomed it into their homes, asking it to take off its shoes before giving it a tour. Before long we made Reddit and Facebook together- some might say we got too buddy-buddy.
I think AI is still dealing with the headline that our culture first understood it with, like the Internet or the platypus. We’re still stuck on understanding the front page of it. AI is still waiting at the door, hat in hand. I’m not here to assuage your discomfort with AI in general- I think a lot of it is probably founded. But I do think there are really neat parts to discover if we allow space for complexity.
We spend a lot of time thinking about the front end of AI, about what it can do for us and how it can do it. Understandable fear about the drivers, doctors, and marketers it may replace. AI is the mammal that lays eggs, and we are transfixed by this. We are surprised every time by the spurs. We don’t stop to think about the back end of AI, that a lot of its “genius” comes from us, from the people that fed it what it is. Absurdity to nuance.
I will be the first to say that I don’t really understand Generative AI, but I do have a vision for where we should go. The disability community can really benefit from a deeper understanding of the back end of AI. Disability communities are siloed and disjointed from their roots. Different and at times conflicting priorities and functional needs lay upon a foundation of disability-specific social bias and medical considerations.
I have worked in Accessibility and Independent Living for over a decade, and despite living and breathing disability, it was years before I actually spoke to someone in the Deaf community. A person with spina bifida has an almost totally different lived experience than someone who is quadriplegic. Disability doesn’t define them but I do think it comes preloaded with a particular set of advocacy organizations and specific support networks.
Your AI should try to understand functionally who you are. I am not advocating for anyone to break HIPAA protocol. I really just want to say that the more your AI understands your functional context- what your limitations are, the environment in which you live, and the things that you want to do, the better the AI will be at giving more specified recommendations and matching you to similar people.
We will become more confident in AI if we know where it is coming from and how it is fed–we need a transparent feedback loop to navigate this. Not just in the way of giving the URLs to certain articles but in the way of saying, “92% of people with similar grip strength and motor abilities to you recommend this product for opening cans.” And then after you used it, “Hey chief, how’d you like that can opener?” People trust what they transparently help build.
People could be connected in a nuanced, ability-first way. and benefit from the things similar people would recommend. This person-centered system would become stronger the more it is authentically used. The system could go beyond the traditional “memory” of a chatbot to formally mediate connections between people and experiences.
We need to stop anxiously looking at it through a slot in our blinds and let the platypus in slowly, telling it to leave its spurs at the door. What if AI could be hyper-relevant, not generic in its answers? What if AI went from connecting people to information, to situationally streamlining the connection between the exact right people?
comments