Artificial Intelligence some questions and answers from a Librarian’s perspective

What is Artificial Intelligence (AI)?

Broadly speaking, AI is intelligence shown by machines, particularly computer programs. It is a field of research in computer science dedicated to developing and studying software and methods that allow machines to perceive their environment and use their intelligence, bolstered by learning to take actions that will allow them to increase its chances of achieving defined goals.

Are any machines or computer programs actually intelligent?

To put it bluntly no, currently all publicly available “AI” programs are basically a type of machine learning programs known as “large language models” (LLMs) that have been trained on massive amounts of text and images, and with set parameters can generate text and images. True machine intelligence has been hypothesized in science fiction but until the AI singularity occurs true Artificial Intelligence will remain in the realm of science fiction.

What is the AI Singularity?

This is the hypothetical point at which AI systems will achieve the ability to surpass human ingenuity and be able to improve itself and soon after this achieve true intelligence and then (possibly) attempt to destroy humanity.

So how are ‘Large Language Models” trained?

The people programming them load vast quantities of text and images into the program to allow the program to analyze these and learn how to write similar works.

Where did they get these vast quantities of text?

They scraped the entirety of the world wide web and downloaded everything they could find to feed the digital beasts they were creating.

And did they pay for all the works not in the public domain that they downloaded?

They did not!

Why not?

To legally acquire all the material still under copyright would have bankrupted them all! To quote (Generative AI) company Anthropic: today’s general-purpose AI tools simply could not exist if AI companies had to pay licences for the material

Why did they not just use materials that are in the public domain?

To quote a spokesman from OpenAI (creator of the LLM program ChatGPT): “copyright today covers virtually every sort of human expression – including blog posts, photographs, forum posts, scraps of software code, and government documents – it would be impossible to train today’s leading AI models without using copyrighted materials. Limiting training data to public domain books and drawings created more than a century ago might yield an interesting experiment, but would not provide AI systems that meet the needs of today’s citizens.”

So they didn’t want to pay to access all copyrighted material to train their programs because it would have been too expensive?

Basically yes!

Is that even legal?

Writers, musicians, artists, other creators and a bunch of publishers and more claim that it isn’t! There is a statement that has been released that has amassed over 40 thousand signatures protesting the use of copyright materials in LLM training: The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works, and must not be permitted.

The creators of LLM AI programs claim that it is fair use to use copyrighted materials to train their programs: we believe that legally copyright law does not forbid training.

So a legal grey area that will have to be fought out in the courts?

mmmmm

OK but is it ethical?

Is it ethical to use a program that has been created using pirated materials in libraries? To quote article 4 of the American Library Association’s Code of Ethics: We respect intellectual property rights and advocate balance between the interests of information users and rights holders. I think as Library Workers we should not be using tools that have been created using methods that may be considered legally, ethically or morally suspect. Most, if not all of us are aware that a majority of authors, musicians and artists struggle to survive on the money they make from what they have created, so to give tacit approval to companies owned by billionaires stealing what they have labored over for years by using Generative AI products so we can push a button to create a generic library lesson plan or library poster that may have one too many fingers or other weird digital glitch is not worth it.

What if I really want to use Generative AI programs at work?

Well I can’t stop you, but I will encourage you to read up on the ethics of using such programs as well as the impact they have on the environment and author incomes. I will put some links to articles below if you would like to start here.

Links:

OpenAI warns copyright crackdown could doom ChatGPT – archived Telegraph article

Explained: Generative AI’s environmental impact – MIT news

Generative AI requires massive amounts of power and water, and the aging U.S. grid can’t handle the load – CNBC

Does the Use of Copyrighted Works to Train AI Qualify as a Fair Use? – Copyright Alliance

Meta Secretly Trained Its AI on a Notorious Piracy Database, Newly Unredacted Court Docs Reveal – Wired article

Authors file a lawsuit against OpenAI for unlawfully ‘ingesting’ their books – The Guardian

UC student explores AI art ethics – University of Cincinnati News

Training Generative AI Models on Copyrighted Works Is Fair Use – Association of Research Librarians

Fair Use Defenses in Disruptive Technology Cases – UCLA Law Review

le chat GPT

Comments are closed.

Post Navigation