Tag: Sentience & Moral Status

  • When Biology Becomes the Computer: Why America Needs a Department of Technology to Govern Biocomputing

    1. The Next Computing Frontier

    Every generation of computing has forced a governance decision on the generation that discovered it. Vacuum tubes gave way to transistors, transistors to integrated circuits, integrated circuits to the networked, GPU-driven systems now training today’s large models. Elsewhere on this site, we have made the case that quantum computing is the next frontier serious enough to demand its own international pledges and treaties, and that artificial intelligence built on silicon is already outrunning the elected institutions meant to govern it. Biocomputing is the frontier after that one, and it is different in kind, not merely in degree.

    Biocomputing does not run on silicon at all. It runs on living tissue — neurons grown in a laboratory, wired into electrodes, and taught to process information the way brains do. That single fact changes everything about how the technology should be governed. A silicon chip that fails is scrapped. A biological system that fails, or that succeeds too well, raises questions no purely technical regulator is built to answer. This is precisely the kind of technological convergence — part computer science, part neuroscience, part bioethics, part national security — that we have argued only a cabinet-level Department of Technology is positioned to govern coherently.

    2. What Is Biocomputing?

    Biocomputing is the use of living biological material, most often lab-grown clusters of human or animal neurons called organoids, to perform computation. Researchers culture stem cells into three-dimensional neural tissue, place that tissue on a multi-electrode array, and use electrical, chemical, or optical signals to train it — much as a silicon neural network is trained, except the network is alive.

    This is no longer a thought experiment. Cortical Labs, an Australian biocomputing company, has demonstrated neuron cultures on electrode arrays that can adapt their behavior in a simplified video game environment, learning from feedback in a way that mimics goal-directed play. A team at Indiana University has combined brain organoids with conventional hardware to perform basic speech-recognition tasks. FinalSpark, a Swiss company, has built a cloud platform that lets researchers run remote experiments on living neural tissue over the internet. The U.S. National Science Foundation launched a program in 2024 specifically to fund this field, and — notably — required every research team to include a bioethicist as a co-principal investigator, evaluated on equal footing with the science itself. Researchers at institutions including Johns Hopkins have organized around the term “organoid intelligence” to describe the broader vision of biological tissue as a computing substrate.

    What exists today are small cultures of tens of thousands of neurons performing narrow, supervised tasks — nothing close to a brain, and nothing close to a mind. What is being actively pursued is scale: larger, longer-lived, more densely networked organoids designed to perform more general computation with a fraction of the energy that silicon AI systems consume. What remains speculative is whether, at some future point, biological complexity of this kind could give rise to something with morally relevant experience. All three of those categories — the real, the researched, and the speculative — matter to this article, and we will keep them separate throughout.

    3. Why Biocomputing Creates a New Governance Challenge

    No single American institution is responsible for biocomputing today, and that is the problem. Stem cell research is overseen by university institutional review boards and the FDA. Basic biocomputing research is funded through the NSF and NIH. Defense applications, if and when they emerge, would fall under the Department of Defense and its research arms. Export controls and international coordination would run through the State Department. Consumer-facing biocomputing platforms answer to no dedicated regulator at all.

    The NSF’s decision to require an ethicist on every funded biocomputing team is a meaningful signal that the federal government already senses this technology is different. But it is a grant condition, not a law. It applies to one funding stream, not to industry, not to universities working outside that program, and not to any foreign laboratory. Ethical review of this kind should not depend on which grant a research team happened to apply for.

    This is a familiar pattern to readers of this site. We have watched the same fragmentation play out in artificial intelligence policy, in quantum computing, and in student data protection: a powerful technology advances across a dozen agencies at once, each with jurisdiction over a slice of it, none with authority over the whole. Biocomputing adds a dimension those technologies did not have. It is not just powerful and dual-use. It is alive, and that raises the ethical stakes of getting the governance structure wrong.

    4. The Need for a Department of Technology

    We have proposed elsewhere on this site that the United States create a cabinet-level Secretary of Technology, nominated by the President and confirmed by the Senate, with counterparts elected at the state, county, and local level. Biocomputing is one of the clearest cases yet for why that office needs to exist.

    A Department of Technology would not replace the FDA’s authority over stem cell safety, or the NSF’s role in funding basic research, or the Department of Defense’s national security mandate. It would do what none of those agencies is chartered to do: look across the whole of biocomputing — research, industry, military application, and international posture — and set standards that apply no matter which laboratory, company, or agency is doing the work.

    This site has also proposed the RMS classification framework — Responsive, Memorable, Sentient — as a clear, legislatively usable way to describe the cognitive complexity of artificial intelligence systems. RMS was built for silicon. Biological systems will need a parallel but distinct classification, because a living neural culture is not a language model, and treating it as one in law would badly understate what is actually at stake. A Department of Technology is the right body to build that framework, in coordination with the NSF, NIH, and the existing bioethics community, rather than leaving each biocomputing lab to define its own terms — which is precisely the ambiguity that biocomputing researchers themselves have flagged as an obstacle to responsible oversight.

    5. Biocomputing Must Not Become a Weapon

    Biocomputing should not become a tool of warfare, and the case for that prohibition is stronger, not weaker, than the case we have made elsewhere on this site against weaponized AI.

    Silicon AI systems are opaque, but they are at least deterministic and auditable in principle — their weights can be inspected, their behavior can be tested against a fixed specification. A living neural culture adapts, reorganizes its own connections, and behaves differently across nominally identical experiments, the same way no two brains develop identically. Building weapons systems, targeting logic, or autonomous decision-making on a substrate that cannot be fully specified or reliably reproduced is not just ethically fraught — it is a genuine escalation risk. A biological targeting system that behaves unpredictably under stress is not a safer weapon; it is a less controllable one.

    There is also a dual-use overlap here that does not exist with silicon AI. The laboratory skills, cell lines, and equipment used to grow and interface with neural organoids sit adjacent to the skills, materials, and equipment used in biological research more broadly, including research that could contribute to biological weapons. A Department of Technology should treat that adjacency as a reason for firmer lines, not looser ones.

    6. A Prohibition on Biocomputing for NBC Warfare

    We have previously proposed that a future Secretary of Technology lead an international treaty prohibiting the use of artificial intelligence in nuclear, biological, and chemical (NBC) warfare. Biocomputing needs its own explicit prohibition within that same framework, because it intersects with NBC weapons development along more paths than conventional AI does.

    A Department of Technology should advocate for a clear, enforceable ban on using biocomputing for:

    • weapons development of any kind, including research, simulation, or design work;
    • autonomous weapons systems and targeting logic;
    • offensive military applications generally;
    • biological weapons research, given the direct overlap in laboratory technique and materials;
    • chemical weapons research and development;
    • nuclear weapons research where biocomputing materially contributes to design, simulation, or optimization;
    • any NBC warfare application, including the optimization, design, targeting, or deployment of nuclear, biological, or chemical weapons; and
    • the deliberate creation of biological computing systems intended to facilitate warfare in any form.

    These prohibitions matter for a reason specific to biocomputing: unlike a line of code, a biological system built for one purpose cannot simply be repurposed by editing a file. The laboratories, cell lines, and expertise built up to weaponize biocomputing would themselves become a standing capability, difficult to dismantle and easy to proliferate. A bright-line prohibition, established early and advocated for internationally, is far more achievable than trying to contain a biological weapons-adjacent capability after it already exists. Modeled on the No First Use Quantum Pledge we have proposed for quantum technologies, we propose a parallel No Weaponization of Biocomputing Pledge: a commitment among signatory nations that living computational systems will never be developed, procured, or deployed for military or warfare purposes, verified through the same kind of transparency and inspection mechanisms we have proposed for quantum governance.

    7. The Sentience Question

    A note on terminology before this section begins: elsewhere on this site, “Sentience” refers to the operating system concept we have proposed for the convergence of AI, robotics, and internetworking. This section uses the word in its ordinary philosophical sense — subjective experience, self-awareness, or another morally significant form of cognition — and the two should not be confused.

    No credible researcher claims that today’s neural organoids are sentient. The cultures now in use contain a small fraction of the neurons in a human brain and perform narrow, supervised tasks. But the field’s own literature is candid that consciousness itself has no settled scientific definition, which means there is no clean test researchers can run to confirm or rule out morally relevant experience as these systems scale in size and complexity. That is not a reason to dismiss the question. It is the reason the question needs a governance answer now, before it becomes an urgent one.

    The ethical problem is not “is this organoid sentient?” It is: what should society do if researchers cannot confidently answer that question at all? Waiting for certainty is not a neutral choice. If a biological system were capable of some form of subjective experience and researchers kept experimenting until they were sure, the harm — if it existed — would already have occurred, repeatedly, before anyone acted on it. A Department of Technology should treat that asymmetry as the starting point for policy, not a philosophical afterthought.

    We propose that a Department of Technology establish sentience safeguards and experimental stop criteria before biocomputing experiments of significant complexity are permitted to proceed, built around six commitments:

    1. classify experiments according to their potential cognitive and biological complexity, not merely their stated research purpose;
    2. subject experiments to progressively stronger oversight as that complexity increases;
    3. continuously monitor high-complexity experiments for predefined indicators of unexpected cognitive capability;
    4. require every high-complexity experiment to have an independently reviewed shutdown protocol before it begins, not after concerns arise;
    5. require immediate suspension when predefined warning thresholds are reached, without waiting for research-team consensus; and
    6. route any evidence of potentially morally significant cognition to independent scientific, ethical, and governmental review, not to the research team alone.

    8. Establishing Experimental Stop Criteria

    The hardest question this article can ask is also the most necessary one: at what point should a biocomputing experiment be stopped if there is credible evidence the biological system may be developing something like sentience?

    There is no simple answer, and any article that pretends otherwise is not being honest with its readers. What can be proposed is a defensible standard, built on precaution rather than proof: credible evidence of potentially morally significant cognition → pause → independent evaluation → heightened safeguards → a determination of whether the experiment may ethically continue. Under this standard, a pause is not an accusation that a system is sentient. It is an acknowledgment that the question is serious enough to warrant a second, independent look before continuing.

    What should trigger that pause? None of the following is proof of sentience on its own. Each is a scientifically credible indicator that would warrant independent investigation, and a Department of Technology should require that every high-complexity biocomputing experiment be monitored for signs including:

    • learning capability beyond what the experimental design anticipated;
    • persistent memory that outlasts the specific training session;
    • adaptive behavior of increasing complexity over time;
    • evidence suggestive of self-modeling;
    • communication or output that cannot be adequately explained by the experimental design;
    • apparent preference or aversion in response to stimuli;
    • persistent, goal-directed behavior that continues across sessions;
    • behavioral responses suggestive of distress or suffering;
    • unexpected integration of information across time; or
    • behavior suggestive of self-preservation.

    The purpose of listing these is not to suggest that any current biocomputing system exhibits them. It is to give researchers, review boards, and a future Department of Technology a concrete, falsifiable checklist — something more actionable than a vague instruction to “proceed carefully.”

    9. The Precautionary Principle

    Underlying all of this is a single ethical principle we believe deserves serious consideration, not as established fact but as a working standard for policy:

    If society creates a biological system capable of potentially experiencing the consequences of our experiments, society assumes a responsibility to protect that system from unnecessary suffering, exploitation, and destruction.

    That principle only does useful work if it is paired with careful distinctions, because these six concepts are not the same thing and collapsing them does real damage in either direction:

    • Biological activity — cells firing, signals propagating — is present in every organoid today and implies nothing about experience.
    • Intelligence — the capacity to solve problems — is present in narrow forms in today’s biocomputing systems and, separately, in plenty of non-conscious processes.
    • Learning — adapting behavior based on feedback — has been demonstrated in neuron cultures and also occurs in systems no one considers conscious.
    • Consciousness — some unified, first-person point of view — remains scientifically undefined even in well-studied animal species, let alone lab-grown tissue.
    • Sentience — the specific capacity to experience sensations such as pleasure or pain — is a narrower and more testable claim than consciousness generally, though still contested.
    • Moral status — whether a system’s interests deserve protection — is an ethical conclusion, not a biological measurement, and does not automatically follow from any of the categories above.

    These distinctions matter because researchers in this field have themselves warned that overstating what current systems can do — describing narrow pattern-learning in a dish as if it were closer to a mind — risks a public backlash that could set back legitimate, ethically supervised research. Understating what future systems might become carries the opposite risk. A Department of Technology should hold both risks in view at once: neither sensationalizing today’s organoids nor dismissing tomorrow’s.

    10. Independent Oversight and Transparency

    Safeguards that exist only inside the research team that stands to benefit from a positive result are not safeguards. A Department of Technology should establish:

    • a national biocomputing oversight board, independent of any single university, company, or funding agency, with authority to pause or terminate high-complexity experiments;
    • mandatory registration and public reporting of high-complexity biocomputing experiments, similar in spirit to clinical trial registries;
    • transparency requirements proportional to experimental complexity, so that low-risk research is not buried in paperwork while high-risk research cannot proceed in the dark; and
    • whistleblower protections for researchers who raise safety or ethical concerns about biocomputing work, extending the protections we have already proposed in our AI Whistleblower Protection Act to this domain specifically.

    11. National and International Standards

    Domestically, a Department of Technology should set a floor that no state, university, or company can undercut — the same role we have proposed it play in AI legislation and data sovereignty. A biocomputing lab should not be able to relocate to a state with weaker oversight and continue high-complexity experiments that would have triggered review elsewhere.

    Internationally, the challenge is harder, because biocomputing research is already global. China issued national ethical guidelines for high-risk stem cell research in 2025; European and Australian labs are active in this field; and biocomputing platforms accessible over the internet mean a researcher anywhere can run experiments on organoids housed in another country entirely. A Department of Technology should treat this as an argument for faster American leadership on international standards, not a reason to wait for consensus that may never arrive.

    12. America’s Opportunity to Lead

    We have argued elsewhere on this site that the United States was late to the table on AI regulation, publishing a national AI Action Plan after other governments had already moved, despite having proposed a clearer framework first. Biocomputing offers a chance to get the sequence right: establish domestic standards and an elected, accountable Department of Technology before — not after — the technology matures into something with military applications or unresolved moral status.

    A Secretary of Technology could use that domestic credibility to advocate internationally for:

    • a global prohibition on military applications of biocomputing;
    • the No Weaponization of Biocomputing Pledge described above, extending the No First Use model we have proposed for quantum technologies;
    • shared international safety standards for high-complexity biocomputing research;
    • transparency and reporting requirements for high-risk experiments, applied across borders;
    • independent oversight mechanisms recognized by treaty rather than left to individual national discretion; and
    • internationally recognized criteria for suspending experiments where credible evidence of potentially sentient biological computing emerges.

    Over time, this could grow into a formal international convention on biocomputing, following the same path we have proposed for our International Treaty on Quantum Intelligence — a dedicated, enforceable agreement rather than a patchwork of national guidelines that stop at each country’s border.

    13. From Technological Innovation to Technological Responsibility

    Nothing in this article argues against biocomputing research. The potential upside is real: dramatically more energy-efficient computation, new insight into learning and memory, and biomedical applications that silicon-based AI cannot offer on its own. The argument is narrower and, we think, harder to dispute — that the same country capable of funding and leading this research should also be capable of governing it, and that the governance should exist before the hardest cases arrive, not after.

    This is the same argument we have made about artificial intelligence, quantum computing, and every other technology covered on this site: the purpose of a Department of Technology is not to slow innovation down. It is to make sure humanity’s ability to create powerful technologies does not outrun humanity’s ability to govern them responsibly. Biocomputing, because it is built from living tissue rather than silicon, is simply the clearest example yet of what happens when that gap is allowed to open.

    14. Conclusion

    Biocomputing is still young. The systems that exist today are small, narrow, and almost certainly nowhere near anything resembling a mind. But the researchers building them are already asking whether their field has a definition of consciousness precise enough to know when that changes — and by their own account, it does not yet. That gap between capability and certainty is exactly where governance needs to arrive first.

    A Department of Technology, with authority spanning research, industry, defense, and international diplomacy, is the institution best positioned to close that gap: to keep biocomputing out of weapons and out of NBC warfare entirely, to build stop criteria before they are needed rather than after, and to lead the international effort to make sure no nation faces these questions alone.

    If we are capable of creating new forms of biological intelligence, shouldn’t we establish the ethical and governmental framework for protecting them before we create something whose moral status we do not understand?