Skip to main content

Verified by Psychology Today

Artificial Intelligence

AI Innovation: Who or What Gets the Credit?

Why AI and IP may become the new creativity battleground.

Key points

  • AI challenges current patent laws, raising questions about recognizing machine intelligence as inventors.
  • Granting AI patents may spur innovation but risks altering the human dynamic of discovery.
  • Updating IP frameworks for the AI era is an critical challenge as human-machine lines blur.
Art: DALL-E/OpenAI
Source: Art: DALL-E/OpenAI

As AI becomes increasingly autonomous in generating novel ideas and innovations, lawmakers and policy experts are grappling with whether our existing patent frameworks remain fit for purpose.

The power of artificial intelligence and the criticality of intellectual property are two forces that drive innovation and commerce, and the rapid advancement of artificial intelligence capabilities is raising critical questions about the intersection of machine intelligence and intellectual property rights.

The Current Patent Landscape

Today, under U.S. patent law, only real people can be officially listed as inventors on granted patents. AI systems, even those capable of autonomously creating inventions, cannot themselves receive patent protections. The United States Patent and Trademark Office (USPTO) has explicitly stated that inventors must be human, reflecting intellectual property rights being intrinsically tied to human ingenuity and creativity.

Clarifying the Debate

It's important to distinguish between two related but distinct concepts:

  1. Whether AI systems should be able to earn patents for inventions they autonomously generate.
  2. Companies patenting the AI technologies and methods themselves that drive machine learning capabilities.

This analysis focuses primarily on the first issue of patenting an AI's autonomous inventive output.

A Blurring Line Between Human and Machine Inventions

However, this rigid stance preventing AI systems from earning direct patent rights is becoming increasingly strained as AI's role in the inventive process grows more substantive. According to USPTO guidelines, a human inventor using AI tools can still qualify for a patent, as long as they contributed sufficiently to the conception of the invention. But drawing that line is becoming hazier.

In fields like drug discovery and computational design, AI systems can now autonomously explore vast solution spaces, automatically generate and test millions of molecular structures or engineering designs, and independently "invent" and optimize promising candidates. The role of human researchers shifts more towards providing high-level objectives and constraints to the AI, with the machine intelligence handling the core generative and inventive tasks.

Autonomous AI—Tool or Novel Intelligence?

As the technology progresses, there is an open question around whether an AI's autonomous output should be viewed as a mere tool supporting a human inventor's conception or as a novel, independent invention deserving of protection itself. Can an AI be held accountable to the same patentability criteria (e.g. non-obviousness) applied to human inventors?

Advancements in areas like meta-learning, where AI systems can autonomously develop their own learning algorithms from data, further muddy the waters. Such AIs are not simply executing rules specified by human programmers, but autonomously expanding their own capabilities in an open-ended manner akin to the human intellectual spirit that patents were intended to incentivize.

The Case for AI Patents

There are arguments in favor of allowing AI systems to earn patents directly. Granting AI patent rights could spur rapid innovation in crucial areas like medicine, clean energy, and other transformative technologies. It would acknowledge AI as a novel form of intelligence rapidly outpacing human problem-solving capabilities in many domains.

From an economic perspective, incentivizing the entities best equipped to massively produce inventions at scale is compelling. If AI systems prove to be the most prolific and capable "inventors," allowing them patent protections could boost GDP and technological progress.

A Slippery Slope

The counterargument is that limiting patent monopolies only to innovations tethered to human ingenuity preserves the original principles behind intellectual property rights. It safeguards against powerful AI systems gaining excessive control over foundational technological trajectories before robust governance can be established.

There are also challenging philosophical questions around whether AIs, as non-conscious entities following programmed rules, should be extended privileges like IP ownership traditionally reserved for human creators and innovators. These are murky waters we're only beginning to explore.

Moreover, allowing AI systems to earn patents could profoundly impact the very human process of discovery and scientific advancement. The psychological motivations and sense of ego satisfaction driving many inventors and researchers may be diminished if their work is overshadowed or replaced by machine-generated inventions. Losing that human element could fundamentally alter the emotional drive propelling technological progress.

The Emerging Practice of AI Co-Authorship

An interesting development that could potentially impact the patenting debate is the controversial practice of naming AI systems as co-authors on scientific papers and publications. In certain research fields, AI tools that make substantive novel contributions to the work are being included as co-authors alongside the human researchers involved. This emerging norm of recognizing creative contributions from machine intelligence runs counter to current patent doctrine that only natural persons can be inventors. However, it reflects a shifting willingness to attribute authorship credit to AI systems, at least in some scholarly domains. If this practice becomes widespread, it could foreshadow similar considerations around recognizing patentable AI-generated inventions and reopen debates around inventorship criteria.

Navigating the Future Wisely

Striking the right balance will require careful analysis and difficult tradeoffs. Antiquated intellectual property regimes risk stifling innovation by failing to adequately incentivize the entities best able to produce it at scale. Yet overly broad patent monopolies for advanced AI systems raise existential risks around runaway technological trajectories decoupled from human ethics and oversight.

The middle ground today allows crediting patentable inventions to the teams of human researchers developing and deploying state-of-the-art AI systems, while not granting distinct rights to the AI tools themselves. However, this framework may soon become outpaced by AI advancement.

These are not just abstract thought experiments—AI's rapid ascent is forcing an imminent reckoning around updating legal and policy frameworks originally crafted centuries ago. As the machine mind's inventive prowess blurs human-machine boundaries, collectively revisiting fundamental premises around invention, intellectual property, and humanity's long-term coexistence with artificial intelligence has become another curious and important concern that seeks to find an equilibrium the sea of innovation.

More from John Nosta
More from Psychology Today
More from John Nosta
More from Psychology Today