Since the 1700s, US copyright law has protected the right of content creators, such as authors and artists, to profit from their work. Society benefits when content creators are motivated to create valuable content. However, recent generative AI systems such as ChatGPT and DALL-E 2 raise questions about whether these and other laws adequately protect the rights of content creators.
A key issue arises from the difference between copying and learning. Copyright law generally requires that you obtain permission before copying another person’s representation of ideas, facts, or similar information. But copyright law does not protect underlying ideas and facts.
And nothing prevents humans from learning from copyrighted material and creating new content of their own, as long as the new material is not substantially similar to the original material.
This also benefits society. Without it, how can writers and artists learn their craft by studying the work of previous masters in the field?
But humans are no longer the only ones who can learn from previous examples and generate their own new content. Today’s generative AI systems can now do it at a speed, scale and cost efficiency far superior to humans.
For example, image generation systems such as DALL-E 2 and Stable Diffusion learned from millions of images on the web. Now, if you ask for a description of the image you want, it will often produce a surprisingly good image of what you describe.
The ability of these new technologies to produce vast amounts of creative content very quickly and cheaply has the potential to provide great value to society. But one question it raises for her is whether a fair share of this value is provided to the original creators of the content used to train these systems.
For example, an artist with a distinctive style may find it difficult to sell new work to people who appreciate that style if that style can be easily replicated automatically. Also, the news publisher, whose content can now be paraphrased by an AI system that generates it without violating copyright law, has seen significant gains from readers who no longer have to click through to visit the publisher’s website. You may lose ad revenue.
So what do we do about this? One promising possibility is that if these large-scale AI systems can process vast amounts of information much faster and at a lower cost than humans can, they could be used for a variety of purposes. to introduce a new legal system based on the premise that adequate legal protection is in place.
Using an analogy to copyright law, we can call this new legal system a “learning law.” Just as copyright law governs the right to copy content, Learning Lights Act governs the right to have automated systems learn from the material.
Given such laws, one of the key questions is how content creators can invoke protections under the Rights to Learning Act. The easiest way to do this (and protect previously available material) is to automatically apply learning rights protection to all copyrighted material.
The law will provide details about what specific types of AI systems and learning will be covered. Then, to legally learn from this material, the operator of a generative AI system would have to license that right from the owner of the original material.
Another approach is that original content creators should explicitly invoke learning rights protection, such as by posting a learning rights notice (similar to today’s copyright notices). This will benefit owners of generative AI systems as more content will be made available for learning for free, but it will also benefit content creators using Learnright to make their content explicit. It will impose a considerable burden of protection on
Regardless of how the Learning Light Act is specifically defined, this approach will allow society to benefit from the power of generative AI to create new content while providing sufficient incentive for humans to continue creating content. may be utilized.
This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., publisher of the Bloomberg Act and Bloomberg Tax, or its owners.
Author information
Thomas Malone is the Patrick J. McGovern Professor of Management at the MIT Sloan School of Management and the founding director of the MIT Center for Collective Intelligence.
Please write to us: Author guidelines
