I think a lot about the epistemics of what has been branded as “AI”. These tools, if they can be called that, are epistemic annihilators and yet have been carefully constructed and presented in ways that have many people treating them as epistemic agents. In my own studies on epistemics, I look particularly at knowledge-as-authority and it is…. frustrating that these stochastic parrots are given so much credence as experts on anything, much less everything.
I know that there are reasons for this. As a journalist who runs a newsroom, I understand the power of media, especially mainstream outlets and how that power shapes allowed and allowable understandings of the world. The way a lot of outlets have covered these technologies involves a lot of deference to what the tech companies building them say about them.
I think this is true of both fiction and non-fiction media. I think that science fiction in all its forms have primed our cultural understanding of these tools in such a way as to make us more likely to think they are intelligent based largely on a misunderstanding stemming from their name and how human stories have historically represented things named the same thing: AI.
Fiction contains an interesting foil to what Silicon Valley is building: Isaac Asimov’s three laws of robotics.
“
- First, “A robot may not injure a human being or, through inaction, allow a human being to come to harm”.
- Second, “A robot must obey orders given to it by human beings except where such orders would conflict with the First Law”.
- Third: “A robot must protect its own existence a long as such protection does not conflict with the First or Second Law”.
“
There is a tension at play in our real world’s build-out of robotic agents, such as it is. I think about Ian Winchester, who wrote in The Journal of Educational Thought about AI and these kinds of bots and how they get in the way of his work while purportedly being there to assist in the work.
“ And the central question becomes whether or not “human being” or “corporate person” has primacy in the laws governing robots. ”
Winchester’s quote from above was published in 2023, and less than a year from when ChatGPT first went live for the public.
That tension I mentioned is revealed when you look at how Winchester framed it:
“ Among the examples with which I began is that the Editorial Assistant function is now heavily in the hands of the robotic Editorial Manager, and the dominant partner is this robot, not the human being. Indeed, one might say that the former editorial tasks of the Editorial Assistant have largely been replaced by the robot. In that sense the robot is taking away a human job. Robots can and do take human jobs thereby putting human out of work. Indeed this has been what our mechanical and electrical devices have been doing since the Industrial Revolution.
And robots apparently are now capable of helping some humans cheat their way through school and university, ChatGPT being the most recent prominent example. Furthermore in many respects robots can be developed that have capabilities beyond that of any human, especially with respect to the retention of data and its use or the rapid identification of things. One example with respect to human activities is that of being a contestant on the television program Jeopardy where IBM’s Watson was successful in defeating the cleverest of humans at the task of answering trivial questions of great obscurity in a given the answer determine the question format due to the extensive data bank possed by Watson for that purpose. ”
There’s something unquantifiable about human experience. I’d actually argue there’s more than one thing about our lives and our struggles and our joys that can’t be measured quantitatively. That’s why my past studies always incorporated at least some qualitative data, because numbers on a spreadsheet alone can’t… capture all that we are.
I believe there is something insidious about the offloading of not just cognitive ability but trust and conversation to these robots.
“We see a future where intelligence is a utility like electricity or water and people buy it from us on a meter” – OpenAI CEO Sam Altman, speaking to a crowd gathered by BlackRock.
Intelligence- human intelligence- can’t be metered. It can be cultivated or squandered but I’ve always believed anyone can learn and do intellectual work. Some may struggle with different forms of it more than others, but the potential to grow and learn is there, and always has been.
For all their talk of alignment and safety, I deeply believe the people building these tools from Anthropic to OpenAI to the rest are the ones who are misaligned and unsafe. There’s something deeply insidious about how they are building something they believe will end humanity in at least one of multiple potential ways and are doing it anyways because it benefits their monetary interests.
Now these tools that have broken the economic balance of consumer electronics and kept electricity prices high even as renewables progressively lower the price of electricity overall are being built to make us believe we need them and “on a meter” like that very electricity to do intellectual and intelligent work? I refuse the notion.
I think of Adam Savage’s somewhat famous quote, “I reject your reality and substitute my own.” Savage is a builder and, famously, a Mythbuster (like from the TV show) who’s also known as a tech adopter. Yet, he too once refused the notion that AI could do anything creative.
I reject the reality that OpenAI and Anthropic are attempting to build around me and will substitute my own by building a different one.
Do you? Will you?
If so, what does rejecting the enclosed reality of datacenters everywhere and intelligence on a meter look like to you?
For me, it looks like building a space for human art and human intelligence by carving out boundaries that prevent slop from entering.
In my work I refuse not only the use of “AI” to write or think, but I also enforce that if you’re going to work with me you must also not use AI to write for you.
It’s kind of like the paradox of tolerance. If we tolerate these tools eventually the only space for intellectual work will be fully colonized by these tools.
Resist.
This article reflects the author’s own views without aiming to represent GCAS’s position as a whole.