Search the KHIT Blog

Thursday, December 8, 2022

"Malign Technologies" update: AI Natural Language Generation (NLG)

Who might have copyright ownership claims to AI-generated human-readable text?

I posted this today on Facebook. From an interesting article in The Atlantic:
“The world of generative AI is progressing furiously. Last week, OpenAI released an advanced chatbot named ChatGPT that has spawned a new wave of marveling and hand-wringing, plus an upgrade to GPT-3 that allows for complex rhyming poetry; Google previewed new applications last month that will allow people to describe concepts in text and see them rendered as images; and the creative-AI firm Jasper received a $1.5 billion valuation in October…”
OK, question for all of my musician/songwriter friends. If you use the free updated GPT-3 to “write” song lyrics by proxy (i.e., you type in a theme—“Ooooh, baby, you’re gone, and my heart is broken, whah, whah, whah…”—and out pops a Lizzy McAlpine-worthy heartthrob lament), well, who “owns” the lyrics copyright? OpenAI? (But, wait! I thought they are “open source?”)

No one will give a shit unless and until the AI-spawned song becomes a lucrative chart success— after which every music biz IP lawyer in LA and Nashville will be aggressively elbowing each other aside in the sprint to the courthouse to file claims.

Count on it.
BTW: I am a long-time songwriter, of durably nil repute (or $$$). As my wife adroitly put it, I was Quixotically "working in the not-for-profit sector." I may have to screw around with GPT-3, just for grins.
  BobbyG, 1981, "The Once & Future Fool"
While living in Knoxville beginning in the late 1970's I was a member of the East TN Chapter of the Nashville Songwriters' Association. But, once belatedly in college at UTK, where I learned the statistical/economic principle of "expected value" (probability x payoff), it became clear that the average estimated prospective "value" of your lovingly-crafted song was in "basis points" (hundreths of a cent) at best.

I changed careers in 1986, and went first into environmental radiation laboratory science. The old joke: "How do you become a millionaire in the music business? Start out with at least two million dollars."

I opened an account at OpenAI to explore this tech.
It's certainly fast. But my initial use thus far is just Q&A.

Now, the topical focus of The Atlantic piece is "authenticity" of student prose writing in the academic environment.
Essay generation is neither theoretical nor futuristic at this point. In May, a student in New Zealand confessed to using AI to write their papers, justifying it as a tool like Grammarly or spell-check: “I have the knowledge, I have the lived experience, I’m a good student, I go to all the tutorials and I go to all the lectures and I read everything we have to read but I kind of felt I was being penalised because I don’t write eloquently and I didn’t feel that was right,” they told a student paper in Christchurch. They don’t feel like they’re cheating, because the student guidelines at their university state only that you’re not allowed to get somebody else to do your work for you. GPT-3 isn’t “somebody else”—it’s a program.
Yeah. But apropos of my opening scenario, one can quickly come up with a number of broader questions. When big money is at stake (e.g., my recent prior FTX posts), all manner of problems surface—NLG "plagiarism (?)" aside.

My priority interest in AI has principally been about stuff like this:

NLU: Natural Language Understanding.

I asked OpenAI ChatGPT the following:
Can artificial intelligence analyze a prose argument?

Yes, artificial intelligence (AI) can be trained to analyze prose arguments to some extent. However, the extent to which AI can effectively analyze prose arguments will depend on various factors, such as the quality and amount of training data the AI has been given, the specific AI model being used, and the complexity and nuances of the argument being analyzed.
I'll have to see if I can provide it an "argument," and instruct it to analyze and evaluate it. Skeptical. but, we'll see.

See some of my prior riffs on "AI" topics more generally here.


More from The Atlantic piece.
A chasm has existed between humanists and technologists for a long time. In the 1950s, C. P. Snow gave his famous lecture, later the essay “The Two Cultures,” describing the humanistic and scientific communities as tribes losing contact with each other. “Literary intellectuals at one pole—at the other scientists,” Snow wrote. “Between the two a gulf of mutual incomprehension—sometimes (particularly among the young) hostility and dislike, but most of all lack of understanding. They have a curious distorted image of each other.” Snow’s argument was a plea for a kind of intellectual cosmopolitanism: Literary people were missing the essential insights of the laws of thermodynamics, and scientific people were ignoring the glories of Shakespeare and Dickens.

The rupture that Snow identified has only deepened. In the modern tech world, the value of a humanistic education shows up in evidence of its absence. Sam Bankman-Fried, the disgraced founder of the crypto exchange FTX who recently lost his $16 billion fortune in a few days, is a famously proud illiterate. “I would never read a book,” he once told an interviewer. “I don’t want to say no book is ever worth reading, but I actually do believe something pretty close to that.”…
Ahhh... C.P. Snow. I've riffed on him before here.

Also, the best place to hide $16 billion from our liitle a-literate Genius pal Sam Bankman-Fried is apparently inside a book. 
"The Beast."
Scott Singeisen, a “Professor of architecture” posts a Twitter thread in reaction to The Atlantic article:
Write a well argued response to Stephen Marche’s recent essay in @TheAtlantic that AI will make the college essay obsolete

Stephen Marche's essay in the Atlantic argues that AI will make the college essay obsolete. However, this claim is not well-supported and is ultimately not convincing.

First of all, it is important to note that the college essay serves a specific purpose in the admissions process. It allows admissions officers to get a sense of the applicant's personality, writing abilities, and potential fit for the institution. AI, on the other hand, is not capable of providing this type of insight. While AI may be able to analyze data and generate texts, it cannot replicate the human experience and provide the same level of depth and understanding that a college essay can.

Furthermore, the use of AI in the admissions process raises concerns about fairness and bias. AI systems are only as good as the data they are trained on, and if the training data is [sic] biased, the AI will also be biased. This could lead to admissions decisions that are unfair or that do not accurately reflect an applicant's abilities and potential.

Additionally, the use of AI in the admissions process could potentially widen the gap between students from different backgrounds. Students from disadvantaged backgrounds may not have access to the same resources and support as their more privileged peers, leading to a disadvantage in the admissions process.

In conclusion, while AI may have the potential to assist in the admissions process, it is not capable of replacing the college essay. The college essay serves a specific purpose in the admissions process and is an important tool for admissions officers to evaluate applicants. The use of AI also raises concerns about fairness and bias, and could potentially widen the gap between students from different backgrounds.
I'm gonna have to play with some of this stuff. 
I'm lucky. I never needed AI NLG "help" when it came to writing—prose or song lyrics. Likely has something to do with the 2-3 books a week (plus all of my periodicals) I've studied across my-now 55 years of "adult" life since turning 21. Ya think?

 Stephen Marche:

"[N]atural-language processing is going to force engineers and humanists together. They are going to need each other despite everything. Computer scientists will require basic, systematic education in general humanism: The philosophy of language, sociology, history, and ethics are not amusing questions of theoretical speculation anymore. They will be essential in determining the ethical and creative use of chatbots, to take only an obvious example…"

Yeah, And, don't forget the lawyers. Never overlook the lawyers, lest you come to rue the day.



My latest login. The popularity of ChatGPT is melting their servers. 

No comments:

Post a Comment