Even if the core part of your business is something other than art, music, or writing, you are still impacted by the changes in computer generated creative works . Does your business have a website? Does it advertise? Do you need to write proportional materials, job descriptions, memos, or other documents? If so, you should be paying attention to the massive changes occurring in artificial intelligence being able to mimic what used to be core areas of human creativity.
Someone with no art training can now tell a computer to produce very artistic images.
History of Computers Replacing Human Tasks
The digital age began many decades ago, and people have generally become comfortable with computers being able to duplicate human thought when it comes to routine or complicated tasks. People generally do not think twice about software that runs arithmetic operations, organizes spreadsheets, compresses data, or does any number of precise tasks. We forget that even the humble calculator was once revolutionary and disruptive. Indeed, the very word “computer” originally referred to human beings employed specifically to calculate the answers to mathematical questions by hand. It was only later applied to the digital machines that put them all out of work.
Since roughly the beginning of the 21st century, the scope of what a computer could do has been getting broader and broader. In the old days, in order to automate a task, you would need to figure out how to give the computer specific instructions on how to move numbers in its memory to accomplish that task. Changes in programming techniques, from the procedural steps of BASIC to the object-oriented languages like C++ certainly helped. But, many problems still remained intractable for this kind of solution.
Enter machine learning. Very roughly speaking, this consists of setting up software that writes itself. Instead of figuring out the precise instructions a computer should execute to estimate the price of a stock, for example, the programmer uses a large dataset of known stocks and related information and then asks the program to find a “best-fit” equation that relates the information. If you’ve ever done data analysis or statistics, think of this as advanced graph fitting. However, instead of coming up with a neat mathematical equation, the software instead typically utilizes a neural network which, while simply code, conceptually mimics the organization of the human brain.
A Revolution in Computers Doing Creative Work
Behind the scenes, machine learning has been quietly changing the world for decades now. Your cell phone can give you directions to any address and account for expected traffic. Your e-mail spam filter is able to identify which letters are probably trash. Advertisers and entertainers online determine what content you are likely to want to see. Computers are now undisputed masters of Chess. But, like a frog in a boiling pot, it seems like the public at large has been only dimly aware of how the world is changing as the kind of problems computers can solve have grown.
That all changed in the fall of 2022. Around that time, two major pieces of artificial intelligence software exploded into the popular conscience: large language models and image generating software. The large language models were able to produce shockingly good, human-like responses to questions. The image generating software, in turn, generates shockingly good photos or pieces of artwork. Work is also being done on systems that can generate music, write software, or even try to provide legal or medical advice (the latter applications having questionable legality). These developments have captured not only the popular imagination, but also the attention of the business community. Capital is flowing into start ups or major companies rushing to use these technologies.
These developments pose many important questions. Some are philosophical – what does it mean for the nature of creativity when a computer can match or exceed the talent of a human? Others are existential – how close are we to producing artificial intelligence that is so capable that it could damage our infrastructure or lifestyles in pursuit of its goals?
Practical Legal Problems Posed by Artificial Intelligence
As a business lawyer, the questions that I am grappling with are more practical. To begin with, the law does not really have a good framework for understanding how to treat technology that mimics human creativity. There are already at least three major lawsuits pending that allege the way that some of these models are trained infringes on licenses or copyright. The U.S. Copyright office has announced that it will not allow computer generated images to be copyrighted. It remains to be seen how these cases will play out, but anyone looking to get into this field will need to think well ahead of the legal developments. A start-up using an art generator might, for example, want to source its models from areas that steer clear of copyrighted work entirely. It may want to explore using trademark, patent, click-wrap contracts, copyright on the characters and other elements of the work, or other mechanisms to ensure it can profit from its work.
Another major practical question is the displacement of workers. While it seems unlikely that artificial intelligence can replace very skilled or well known authors, artists, lawyers, or journalists anytime soon, we are already witnessing devastation in what I would call the “functional” tiers of these industries. If you want to hire an artist to draw your on-line avatar or book cover, now it’s cheaper and easier to simply have hundreds of pieces generated by software and then choose the best. If you want text for your website to flesh it out, instead of hiring a writer you can simply generate some reasonably good text using a large language model and edit it. In other words, if a business is looking for art or text more or less just to have it, and is not overly concerned with exactly how it turns out, it will turn to artificial intelligence. This is going to put people out of work. That has already caused some lawsuits to filed and more are sure to come.
On a related note, since these technologies allow people to generate stories, artwork, music, and so on without going through the usual long period of training and effort to learn the skills themselves, it is inevitably leading to a greater volume of creative work. I have heard that at least one major science fiction publisher has stopped accepting submissions because of the intractable volume of work being submitted now that people can write using artificial intelligence.
Simply put, this is a period of great instability for creative employees and great opportunity for business owners. The precise legal and functional details remain to be seen, and I am fascinated to learn what happens next.