When I talk to my classes about why they all have to take Comp1, I emphasize that writing is a product of thinking. We meet for class time so they have a break from the busyness of the world to think and reflect in ways you may not have time for during work, family time, or social hour. Thinking is real work, and we get blocked and stumble around and draw blanks. We don't even know where thoughts really come from. We may as well invoke the old muses.
Perhaps that's why generative AI is so startling-- it can write, draw, and even code with such speed. It seems able to think. It's not flawless, but that's beside the point when it can generate dozens of ideas while we come up with zero. We remind ourselves: The illusion that these programs are thinking, and even sentient, is only screen-deep.
Or is it an illusion? Are we merely pawns in a chatbot's chess game? Who's playing who? After last week's drama, I feared it would be time to address the dry stuff like copyright or ownership (and I will do so before long...). But generative AI generates headlines, and really, you can't make this stuff up...especially when it's about a chatbot that makes stuff up. These tools are powerful because they seem to embody...someone. With a mission. Some fear the ghost in the machine (a philosophical concept, as well as a lively album by The Police). Maybe the ghost is now haunting us. Maybe we are the ghost.
"Sci-Fi Publishers Are Upset Over Heaps of Unwanted AI-Generated Pitches," Armani Syed for Time, February 23, 2023
There's nothing futuristic about Spam. But there might be something futuristic about a sci-fi magazine getting bombarded with spammy pitches for sci-fi short stories. Talk about the rhetoric of failure...
"ChatGPT launches boom in AI-written e-books on Amazon," Greg Bensinger for Reuters, February 21, 2023
As of this writing, dozens of books co-authored by ChatGPT are available on Amazon. I mean....they credited their sources, right? And here we were worried about AI being used for plagiarism. Lest you think every little thing it does is magic, sift through the list. Most of them look pretty abysmal.
"AI's rise generates new job title: Prompt engineer," Peter Allen Clark for Axios, February 22, 2023
For those whose writing ambitions run more along the lines of tossing out an occasional tweet, you might be perfect for this job. Not sure what it means? Right up there with the "chief purpose officer," which made my list of all-time awesome job titles, this skill set-- or maybe a full-fledged job-- would be great for someone who knows how to pare down too much information into an elegant twist of phrase.
"ChatGPT listed as author on research papers: many scientists disapprove," Chris Stokel-Walker for Nature, January 18, 2023
It's always fun to hear academics pontificate over the definitions of terms, and this is no exception. Among other details in the debate, I am thrilled that one of the preprint servers is named HAL. Can ChatGPT rack up authorship credits in the ivory tower? Not likely, given the reasons in this piece. It would be nice to see academia rehumanize itself.
"ChatGPT, Bing Chat and the AI ghost in the machine," Gary Grossman for VentureBeat, February 21, 2023
This piece reflects on the mayhem generated by the chatbot conversations that went off the rails last week (and which I also included in last week's post). There's quite a bit to ponder in this piece-- hidden layers, theory of mind, the preponderance of archetypes-- plus the description of these chatbots as "confused models" who get overwhelmed. Us too, ChatGPT. Us too. Perhaps with a little time, you will find light in the darkness...you will make some sense of this...
Comments