Intractable problems, creative solutions

When Christopher Columbus returned to Spain in 1493 convinced that he had discovered a western passage to India he was given a jubilant reception. Jealous of the rewards heaped upon the explorer, and the support to undertake another voyage, Spanish courtiers assailed him. In a famous, perhaps apocryphal story, members of the court minimized Columbus' accomplishments during a banquet arguing that anyone could have stumbled into what came to be called the West Indies. Columbus, they alleged, was merely lucky.

Columbus reputedly responded by having a plate of boiled eggs brought to the table and challenged his antagonists to stand an egg on its tip. One after another failed and insisted that the task was impossible. Columbus then took an egg, slightly crushed the tip, and stood it on its end. The courtiers mocked him again asserting that he practiced no greater skill than any of them could have performed, evoking Columbus' agreement, but also his rebuke that the difference lay in the fact that they might have been able to do it, but he had done it.

Columbus' Egg has long been a reference for creativity combined with action. Many discern a problem, some ponder solutions, but few are bold enough to implement remedies.

 History is replete with examples of challenges that burdened civilizations until they were ultimately overcome. The wheel, the printing press, the cotton gin, and the shipping container all began as groundbreaking inventions that came to transform economies and civilizations. Medications and surgical techniques have improved our responses to plagues, illnesses, and once-disabling trauma.

Mathematician and physicist Archimedes said, "Give me a lever long enough, and I shall move the world." While levers may be in plain sight, the vision to define a problem, determine a solution, and possess the courage and willingness to act is often illusory.

As I enter my 23rd year as a professional mediator, I continue to reflect on the challenge of persuading those in conflict to imagine a world in which all disputants gain from integrative solutions. So ingrained in litigation culture is the old saw that the best settlement is one in which all parties feel unhappy, that it is a challenge to convince parties to look beyond distributive bargaining and allow mediators to help them seek a resolution in which everyone benefits. This problem arises in simple two-party disputes and in what present as broad intractable industry-wide dilemmas.

Consider a current challenge in which we presently see only the tip of the litigation iceberg: copyright litigation involving the use of artificial intelligence (AI). At least 35 such lawsuits are pending in the United States. The United States Copyright Office has issued guidance on registering works where AI played a creative role, made registration determinations, faced scrutiny in federal District and appellate Courts, and is now conducting further study and inviting public comment in anticipation of revisions to its guidance and update of its compliance manual. Very smart lawyers and businesses are investing large sums debating liability over the input of material into large language models (LLMs) and the output from AI platforms.

What if some of those resources were diverted to problem solving in litigation or more globally? Copyright holders, academics, and lawyers generally agree upon certain fundamental principles. Copyright law is intended to foster creativity, protect original expression by an author, and provide fair compensation for its use. Yet there can be an inherent tension between these objectives, with a primary objective being to maintain the balance.

Desiring to compensate copyright holders fairly without impeding creativity can present challenges; broadly that is a primary problem posed by artificial intelligence. This challenge is not new. Conventional solutions may be complicated; they present opportunities and afford insight into how the balance has been maintained in the past. They also demonstrate the power of using a bigger "lever" and perhaps a greater vision.

On the input side, parties question whether that which is ingested into particular LLMs is protected by copyright. This invokes fact-specific inquiries concerning the material and the platform. For example, is the material public domain, does it meet the criteria for copyright protection, is it registered, are rights acknowledged, what is the purpose and scope of the platform, and how does it use the input? The specific facts may lead to later litigation, but preliminary answers to such questions impact the next-level inquiry and problem-solving that could reduce the volume of litigation and foster the still nascent industry.

On the output side, some litigation may remain inevitable if protectable expression is "copied," resulting in substantially similar infringing creations. These again invoke fact-specific inquiries. But history may suggest solutions to sidestep at least some prospective litigation.

Potential solutions are limited only by stakeholders' imagination. For instance, some copyright holders have historically permitted certain uses of their creations through licensing agreements. Creative common license repositories exist for a variety of uses, including web page development, music, photographs, art, film, news, books, video games, games, comic books, technology, and education. Many of these resources operate on an open-source basis permitting free use of materials. Other repositories offer such materials on a limited or perpetual license basis at established fees.

The AI community may be apprehensive embarking on a quest to establish a license structure and pricing for reasonable defined uses to "train" LLMs, and for limited rights to generate non-infringing expressive output. But history suggests that these are not impossible tasks. If a license mechanism is not resolved by negotiation, an approach could follow such as that which occurred in the music industry. After significant litigation, certain music royalties became subject to determination by the Copyright Arbitration Royalty Panel and later by the Copyright Royalty Board. The collection of certain music royalties by performance rights societies may provide a useful paradigm for the administration of AI licensing fees, and the establishment of the Copyright Claims Board following the adoption of the CASE Act could permit expeditious and efficient determination of appropriate size disputes.

In the infancy of AI creation and definition, it would take creativity and effort to define categories of use and valuation and agree upon a methodology for assigning royalty rates. Licensing is not the only or necessarily the proper methodology for seeking to resolve these copyright issues. Agreeing upon a licensing approach may favor large well-funded AI platforms which can afford to build their LLMs with acquired content over less well-funded startups. Smaller limited scope AI platforms could launch with open-source materials, however, minimizing initial barriers to entry. This is but one example of an approach and the historical ability of the marketplace to solve what otherwise appears to be an intractable rights-based problem. The risks of leaving determination of rights to litigation has stymied creation in many industries, discouraging cooperation and creative expansion, and the sharing of rights.

Antoine de Saint-Exupery said, "If you want to build a ship, don't drum up people together to collect wood and don't assign them tasks and work, but rather teach them to long for the endless immensity of the sea." The promise of AI has already transfixed the imagination with its endless possibilities. What I often see in mediation is the stagnating effect on those who focus on the collection of wood and the putative task of assembling while the immense sea beckons. Do not lose sight of clients' objectives and the promise for all disputants that await the day after a dispute is resolved.