AI Gatekeepers and the State of the Art
I am not an AI expert. I got into the conversation about AI and art mostly by accident. A few years back, I was contacted about being a producing partner for a nation-wide consultation on AI and the arts by the AI Impact Alliance (Valentine Goddard) and Akoulina Connell and we eventually delivered 13 workshops that touched every province and territory.
I had a lot to learn and still have a lot to learn. The more I came to understand the changes underway, the more it mattered to me and I wanted to go deeper.
I have a real problem with being asked to acquiesce to '“official” ideologies and saw how AI was serving to amplify these ideologies at the expense of other ways of knowing the world. When I’m told that I should choose between A and B - progressive or convervative, big government or small government, moral outrage or ironic condescension - I find myself wondering about C and D and E and F. I imagine this is quite irritating for those around me, but it’s a sincere posture toward the world, and one from which I seem incapable or unwilling to wrestle free.
A recent talk by Noah Levenson for our community members is sitting heavy with me and I wanted to share some distillations from that conversation and to make clearer my and Ferment AI’s intentions.
AI is not a mature or developed space
Like any field (the arts, for example), AI has its own sets of gatekeepers and its own esoteric jargon. Gatekeepers benefit from making AI seem inaccessible and there are a lot of folks getting paid right now to perform one of a very limited number of scripts of what AI is about. The idea that AI either leads to utopian plenty or dystopian ruin diminishes human agency in ways that I don’t find attractive. By reducing the incredible diversity of human ingenuity into an A/B worldview, we do little to help imagine and deliver a better world(s). Coke and Pepsi spent decades trying to will into existence a world where there were only two choices when we are thirsty. The arts can too easily fall into binaries of commercial and non-profit work or high and low art. Beverages or art, the reality is a constellation of choices with new ideas constantly bubbling up. It seems pretty clear who benefits from reductions in complexity.
We are asked to take sides with the threat that failing to take one will mean being seen as an enemy by both. It is much easier to be noticed (and funded) if you make the “right” kind of AI project. Art money goes to drawing attention to the worst excesses and the vast majority of art about AI is critical and didactic. STEM spaces prefer the optimistic version and reward those that play along. In either case, there’s a script they want us to follow.
Noah offered some simple advice. “Don’t do it.” The world needs as many unique perspectives on AI as there are people willing to think about it. Yes, being a certain kind of critic is profitable, but we are better off when other perspectives are brought into relationship that touch on different aspects of human experience. A complex and nuanced point of view translates poorly to Instagram, and so we are left with two tribes firing volleys across a digital divide.
The worst people working in technology want us to think that the barriers are too hard, the skills required too sophisticated for a novice to enter. Noah offered that seemingly complex challenges are always based on simpler concepts that are understandable if we are willing to go deep enough. Noah’s own story is testament to this and his path is hardly the traditional one. He shared that three particularly good books and endless Google searches allow him to create works previously unseen.
As Noah offered in his talk, working with AI right now is like picking up a camera in 1880. No one at the time imagined or believed that photography could become a legitimate means of creative expression. They were wrong.
Give yourself permission
I dislike authoritarianism regardless of its intentions. All of my work has been about disordering the routines that limit our ability to imagine other paths. The artists I work with are concerned with different things - compassion, or voice, or experiences of wonder. Whatever your thing is, if you bring what you do in the rest of your life into AI then the ends feel more achievable. And, as Noah offered, if nobody is doing what you’re doing that doesn’t mean it’s impossible, it just might take a while. Fermentation feels like an increasingly apt metaphor for this process.
Yes, AI presents a number of hard problems. I’m in the early stages of my own exploration and I am experiencing the thrill of constantly learning new things. I have the confidence of the beginner. My ignorance is vast and new knowledge is exciting. The temptation is to jump into other new things to keep that excitement. But because my questions matter to me, I know that I need to keep with it, even when the excitement fades.
Noah described his own experience of moving out of the new and the exciting and into the hard and the lonely. If no one can help you, you’re probably doing a great job and I hope to become at some future point, “an expert in something that no one else cares about”. I’m starting to develop my own insights, both technical and thematic, and I share them here. They are half-baked and liquid. They will remain that way for a while.
I worry that a convergence of forces is removing our ability to experience and appreciate “unofficial” ways of knowing the world. AI contributes to this, particularly when the tools are in the hands of those that benefit most from the “official” ideologies that dominate our economies and our communities. I want to see how AI can complicate and interrupt these official narratives and to make art that celebrates these moments. I have a ways to go.
There are too many people claiming to have privileged knowledge about AI. Being hard doesn’t mean it is inaccessible. The tools will improve as will the availability of resources, and I’d like to be along for the ride.
There are hustlers and there are gatekeepers, but I have been overwhelmed by the generosity of partners and those with deeper experience in AI open to supporting novices like myself. They want people to talk with and are willing to speed up the process of getting us into the conversation.
How would you like to be in the conversation?
This is the free version of the Ferment AI newsletter. Paid subscribers get free access to our Salon series of conversations and to other opportunities that come up through projects and other partnerships. On November 11, the amazing Mi’Jan Celie Tho-Biaz will be hosting a conversation. Upgrade to a paid subscription and join this talk with Mi'Jan Celie Tho-Biaz to reflect on how we gather stories and how our algorithmic systems 'think' about data.
Mi'Jan Celie's practice incorporates small, public oral history interviews that are thoughtfully designed and carefully curated as performative art spaces which then facilitate a larger witnessing. Mi'Jan Celie will share her practice and prompt a collective conversation about how data is "acquired" in AI systems, particularly AI-based art works.