By Adithi Iyer
The idea of artificial intelligence is just seeping into our collective consciousness, but as we watch new developments in the space, the true “new kid on the block” may be a new type of infused human-technology intelligence — one derived from a blob of cells no larger than a grain of rice. These new units of computational prowess are brain organoids, grown in-lab and capable of producing very basic, but real-time, neurological activity. Brain organoids are a specific, and arguably the most interesting, subset of organoid models that are just beginning to enter legal debates.
The primary use case for brain organoids, like other organ models, at this stage is as a tool to study human neurological development. To produce brain organoids in-lab, researchers employ a method of inducing cell samples to revert themselves to a state of “pluripotency”, turning them into stem cells that have the capacity to divide and differentiate into any type of cell. When these “blank canvas” pluripotent stem cells are cultured in three-dimensional media, they self-organize and begin to differentiate into the different type of brain cells and tissue that can generate, on a miniature scale, functional systems that exhibit measurable neural activity.
Making these organ models especially unique in the existing legal and ethical discourse is their potential to actually “think.” “Thinking,” of course, does not refer to the full spectrum of thought that humans can display. We’re not yet at the point of sentient brain organoids, but we can measure their functions, and they do generate data—and sentience is certainly not out of the question down the line. The neuroplasticity of brain organoids offers a bit of a wrinkle, though; they can learn “unsupervised,” learning through doing and improving faculties in basic skills — Cortical Labs demonstrated something akin to this several years ago with their neural cell dishes learning to play the game Pong.
If we weren’t sure about the personhood or even living status of organoids at large, the lines get especially blurred in the case of brain models. The application of these questions to the output of these organoids is a veritable soup of haze, and a fascinating notion.
Harnessing this output — namely, computational, processing, and cognitive function — is a central focus of the nascent field of organoid intelligence (“OI”), also called Brainoware. In a full-circle twist, OI envisions the use of brain organoids to create organic processing units, essentially wiring the neurons within these organoid cultures to respond to electrical inputs as actual neural networked biocomputers. There are several reasons that OI might be revolutionary for both AI and computing—they’re more energy-efficient than the average supercomputer and, if advanced to the point of fully replicating the structure/function of the human brain, can respond better to complex questions and decision-making problems than pure machine learning.
Hybridizing OI and AI, and adding what seems like a “human” component into our current advances, probably asks more questions than it answers. Here are some of those questions for the law, and how we might begin to think about them.
The Best — and Worst — of Brains
Envisioning how brain organoids might entangle themselves with the law doesn’t take a wild imaginative step; many of the questions we might have around brain organoid models are similar to the ones we’re currently grappling with regarding artificial intelligence. Would OI warrant recognition for the work it produces? And is that output protectible? Under current (and quickly-evolving) copyright developments, AI doesn’t meet the “human” requirement for authorship on its own. But AI (and OI) require human input to work, and there may be some wiggle room on AI work protection, either citing AI as a joint author with human operators, or drawing a line at a certain threshold of human control in the AI-generated work as sufficient for copyright protection.
OI adds yet another layer of complexity in these questions, though. If a brain organoid-operated computer generates a creative work, it introduces to the AI analogy a biological component that shares the same genetic instructions that make up someone’s (the tissue donor’s) brain in real life. Do we have to recognize the donor? Could the donor come after works made by “their” brain tissue, and will that affect how we contract around donation in the first place? The “something more human” element that the organoid processor adds to AI debates will be a fascinating arena for legal development.
Of course, where there’s opportunity, there may also be blame. And assigning blame or liability to the errors and misjudgments of OI is similarly less than clear. AI for medicine and healthcare already deals in questions of how to integrate pure machine learning into diagnostic and clinical decision-making, and with the combination of human programming and machine processing/output — and the ‘black box’ of AI computing — it’s unclear where to place liability for AI-facilitated harms. OI ups the ante; we now have multiple new parties involved in the upstream process of obtaining tissue samples, generating organoids, and modulating and manipulating electrical currents in a certain way to yield specific biocomputational results. How, and around whom, should we construct notions of liability when a potential tort case seeks a defendant? These implications extend beyond the healthcare context, but nonetheless have major questions attached that should keep private enterprise especially on their toes with countervailing issues of liability, justice, innovation, and access all in play.
Next Up for Brain Organoids
These legal questions make up something of a thought experiment around brain organoid intelligence as the field is still in its infancy. But brain organoids as used for research and development are moving at full speed ahead, particularly as organoids enjoy increasingly widespread adoption as stand-ins for animal models in clinical trials. The FDA Modernization Act 2.0, passed in late 2022, officially allowed preclinical trials to adopt these animal model alternatives, and brain organoids seem poised to make a difference in eventually offering more accurate predictions of emerging therapy effectiveness without risking patient lives. Organoid intelligence is a different application, and more downstream perhaps than these more immediate uses, but as with many developments in tissue engineering, it’s one to watch.
