All three of my readers know that I’ve been having fun playing with various AI tools: using prompts to “make art”; employing AI-generated voices at murph.ai for song creation (I cannot sing and can prove it); interviewing ChatGPT weekly; and using tools like Writesonic and Poe to “version” various pieces of writing. This is all light fun. I play to learn, and I hope this technology will inform who I am as a writer, teacher, and artist (whether it means using or rejecting the technology itself).
Yesterday, in the Austin American Statesman, reporter Kara Carlson presents the views of Greg Brockman, co-founder of Open AI, for the future of what is currently the world’s largest celebrity: Open AI and ChatGPT. I did not learn anything from this piece, and that is definitely not Carlson’s fault; it is Brockman’s. The result is another case of our language lagging behind the speed of invention. The creators themselves frequently do not know how to talk about their creations. When they do, the result is rhetoric that is meant to “sound cool” but ends up being deeply unappealing, unsettling, and unimaginative.
Some examples appear below, and you can decide for yourself if this sounds like someone who has a concrete vision, or idea, for what goals exist beyond the act of creation itself. (Yes, insert Frankenstein references here as needed.)
When talking about how AI will “revolutionize everything,” Brockman offers, “We’re clearly moving to a world where (the internet) is alive. You can talk to it, and it understands you and helps you.” Now, I can see his point that our existing web might be viewed as a massive, static form of Wikipedia, but it is his jolt of lightning that brings it all to life. Alive! It’s alive!
But what’s missing? Partially, it is an answer to the question, “Why is that good? Why do we want that? Do we not want this from other people, thus community?” What’s vital is that Brockman is not making the point that that internet’s information itself will change, only that the interface will be more dynamic.
Okay, so let’s game that out: think of all the internet’s failed promises: the disinformation, division, platforms for hate, deep isolation, and rising inequality from what was to be the world’s great equalizer. Yes, let’s supercharge all of those flaws without talking about how we are going to rectify them! Look, I know the internet is powerful and has done significant good. But I also was there at the beginning, learning basic HTML tags when the great promise was that equality and democracy had just received their ultimate shield. Nothing, especially authority (and thus authoritarians) could be centralized enough to ever maintain power. Oops.
Does Brockman try to address this at all? Sort of. There is a weird moment in this article involving Elon Musk (surprise!), where Musk resigned from the Open AI board and plans to develop his own “anti-woke” AI software–what does that even mean? Generally, when people like Elon Musk indicate something should be “anti-woke,” they mean racist and techno-aryan–welcome to our new Crypto Libertarian Paradise! If you are Kevin Brockman, you should react overwhelmingly negatively to such a stupid objective for what is supposed to be a powerful tool. Instead, we get this this watered-down soup:
Brockman acknowledged the criticism, saying the company is not perfect and is working toward evolving the tools.
“It was a failure on our part. We were not fast enough to address biases in ChatGPT. We did not intend them to be there. But our goal really was to have a system that would be egalitarian but treat all the mainstream sites equally, and we actually have a lot of improvements on this over the past month.”
The ethics of artificial intelligence is something Brockman thinks of often, and he sees AI as something all humanity will need to help work on.
“We have a team that works really hard on these problems,” he said. “I think we’re all very aligned in terms of trying to make this technology more trustworthy and usable.”
Kara Carlson, Austin-American Statesman, 3/11/23
It is not a failure when you don’t clear Elon Musk’s bar for anti-wokeness; it means you are moving in the right direction. And why should all “mainstream sites” be treated equally? Are they equal? (See: mainstream site vs. Dominion lawsuit for a quick answer.) And here’s a hint: eliminating bias is a myth. In fact, it’s bad. Look at our “balanced media.” Anti-bias birthed the paradigm of “both sides” (whatever that is) to replace the value we call “truth.”
Then we get to the really good stuff, and by this I mean where people who talk about the future have no idea what they are talking about, where artificial spagetti is thrown at non-intelligent walls:
Brockman said what AI will look like in 2050 is “unimaginable.” He hopes one day Dall-E could be used to make your dreams into art by hooking people up to dream interfaces.
Kara Carlson, Austin-American Statesman, 3/11/23
Why would someone “hope” for this? What is the benefit? Why is it art? Is thought automatically art? You know who is better at thinking about these questions than the creators themselves? Flesh-and-blood writers. I highly suggest Richard Powers’ work (all of it), but his latest novel Bewilderment is a great place to start. Why? Because the novel is centered around an experiment where technology (neurofeedback) is used to help a father’s emotionally troubled son. The technology is powerful and transformative…until it becomes out of reach for those who didn’t create it. You know, like insulin, healthcare, food, or…the internet. Think of Flowers for Algernon for our contemporary world.
Mr. Brockman is not finished. As is common, when there hasn’t been much thought about the future of one’s creation, “entertainment” becomes the catch-all goal:
“If you think about today, where everyone watches the same TV show, and maybe people are on the last season of ‘Game of Thrones,’ ” he said. “But imagine if you could ask your AI to make it go a different way, and maybe even put yourself in there as a main character or something and having interactive experiences.”
Kara Carlson, Austin-American Statesman, 3/11/23
Yes, if only we had such “or somethings.” You know, like a TV show called The Last of Us that is based on this superbly immersive, interactive, individualized technology called…games. Hopefully, some day, such things will be possible.
I agree. This technology will be transformative, and it will be so at the macro-level in a way that is wholly unorginal when compared to other transformative technologies: we will throw it out into the world, consequences be damned, and instead of aiming for specific goals (say, shooting for the moon), all of the non-creators will once again be the custodians sweeping up unintended consequence after unintended consequence.
Based on the language of Mr. Brockman himself, even he is bouncing in the wake of his creation.
In the end, we cannot just complain–we must find ways to meaningfully engage (I’m casting about for such ways), so that we don’t have a repeat of the last great tech pantheon, where wealth equals expertise (Bill Gates is fixing education! All of it!). We don’t need to democratize access to something like Open AI; we need to democratize expertize in its creation and use.