Who Owns a Living Computer? Consent, Cell Lines & HeLa

A living computer built from donated human cells is usually owned by nobody outright. It is a stack of separate claims: the donor’s consent rights, the repository’s cell line license, the company’s hardware, software and inventions, and the customer’s data. The HeLa story shows what happens when the person at the bottom of that stack is left out.

Key takeaways

A living computer is built in steps (donor, cell line, organoid, device, service), and different parties hold different claims at each one. Ownership is mostly settled by contracts, not by a single property right.
Henrietta Lacks’s cells were taken in 1951 without her knowledge. Her family gained a say over HeLa genome data only in 2013, and settled a lawsuit with Thermo Fisher Scientific in 2023 on confidential terms.
Moore v. Regents of the University of California (1990) rejected a patient’s property-based claim to his excised cells but recognized a duty to disclose research and financial interests. It still shapes US thinking.
Most existing stem cell lines were collected under broad consent that did not mention computing. Whether that is ethically enough is a live debate.
Donor genomes are potentially identifiable, so “anonymous” lines are not a complete privacy answer.
Responsible companies can act now: provenance tracking, re-consent where feasible, genomic data protection, benefit sharing and public transparency reports.

This article is general discussion of ethics and policy, not legal advice.

What is a “living computer,” and why is ownership a hard question?

A living computer, in the sense used in biocomputing, is a system that uses living neurons as part of its computing substrate. Today that usually means a brain organoid, a small three-dimensional cluster of lab-grown neural tissue, sitting on a microelectrode array that stimulates it and records its electrical activity. The field calls this organoid intelligence, and our overview of what organoid intelligence is covers the science.

Ownership is hard because the most important ingredient came from a person. An organoid is made from someone’s skin or blood cells, carrying their complete genome, reprogrammed and grown into neural tissue that can be shipped worldwide. The donor may never know where it went.

Our broader article on the ethics of organoid intelligence surveys consent, consciousness and oversight together. This piece goes deeper on one question: who has a legitimate claim to a living computer, and what do they owe each other?

Three related questions sit next door and get their own treatment. Whether organoids could ever have experiences is covered in can brain organoids become conscious?. Whether such systems could have interests of their own is the subject of the moral status of biological computers. And what happens at the end of a culture’s life is discussed in whether it is wrong to switch off a biological computer. Here we stay with the humans in the chain.

Who claims what, from donor to computing service?

The easiest way to see the ownership problem is to follow the material. Most organoids used in research start as induced pluripotent stem cells (iPSCs): adult cells reprogrammed into a stem-like state that can become almost any cell type. Our guide to how brain organoids are grown explains the biology.

Step	What exists	Who typically makes a claim	What the claim usually rests on	Main ethical tension
1. Donation	A skin biopsy or blood sample	The donor (and, in a weaker sense, their genetic relatives)	Informed consent, privacy law, research regulations	Did the donor understand and agree to the uses that follow?
2. Cell line	An iPSC line, banked and characterized	The collecting institution or repository	Material transfer agreements, licenses, institutional policy	Broad consent stretched across decades of unforeseen uses
3. Differentiation	Neural progenitors and brain organoids	The lab or company that grows them	Its own labor, know-how and any patented methods	Whether value added by the lab erases the donor’s interest
4. Device	Organoid on an electrode array with control software	The hardware and software developer	Patents, copyright, trade secrets	Proprietary systems can hide where the cells came from
5. Computing service	Remote access, experiments, trained models, datasets	The service operator and its customers	Terms of service, data licenses	Downstream users rarely see the consent behind the tissue
6. Outputs	Recordings, results, publications, products	Customers, researchers, publishers	Contracts, copyright, data ownership terms	Genomic and neural data may still be linked to the donor

Two features of this chain matter most.

First, claims are layered rather than exclusive. A repository can license a line while the donor keeps a right to stop future use. A company can own its software while the cells inside its devices came from someone else.

Second, the donor’s leverage is greatest at step 1 and shrinks at every step after it. Once a line is immortal and distributed, the donor depends on everyone downstream honoring the terms they agreed to, often without knowing who those people are. That asymmetry is the heart of the HeLa story.

What happened with Henrietta Lacks and the HeLa cells?

The cells taken in 1951

In 1951, Henrietta Lacks, a Black woman being treated for cervical cancer at Johns Hopkins, had cells removed from her tumor without her knowledge or permission. NPR and the Associated Press both reported these facts in their 2023 coverage. The cells grew in culture indefinitely, something researchers had struggled to achieve with human cells, and became the HeLa line. Lacks died at age 31.

HeLa cells became one of the most widely used tools in biology. NPR’s reporting notes their use in research on the human genome and in developing polio and COVID-19 vaccines. Johns Hopkins has said that collecting and using the cells was legal in the 1950s, that it would not happen today without the patient’s consent, and that the institution never profited from them.

Her story later reached a wide audience through the book and film The Immortal Life of Henrietta Lacks.

The 2013 genome controversy and NIH agreement

In March 2013, a team at the European Molecular Biology Laboratory published a HeLa genome sequence (Landry et al., G3). Members of the Lacks family had not been consulted, and the publication raised obvious concerns: a cancer genome from Henrietta Lacks also carries information about her living descendants. The data was withdrawn from public access while the issue was addressed.

In August 2013, Nature reported that the US National Institutes of Health had reached an agreement with the family. Full HeLa genome data would be placed under controlled access, and researchers would apply to use it, with requests reviewed by a working group that included members of the Lacks family. NIH leaders Kathy Hudson and Francis Collins described the arrangement in a Nature commentary the same week. A second HeLa genome study (Adey et al., Nature) was published under the new terms.

The agreement was significant, and also limited. It gave the family a voice over genomic data. It did not give them ownership of the cells, a share of past or future revenue, or control over the countless HeLa cultures already in labs worldwide.

The 2023 settlement with Thermo Fisher Scientific

In 2021, the Lacks estate sued Thermo Fisher Scientific, alleging that the company kept commercializing HeLa cells long after their origin was well known. The company argued the claims were filed too late under the statute of limitations. On August 1, 2023, the family announced a settlement. The Associated Press and NPR both reported that the terms were confidential.

Because the terms are confidential, the settlement creates no public legal rule. Its significance is social: a negotiated outcome with a commercial user of the line.

What HeLa teaches biocomputing

Modern lines are collected with consent, under ethics review, so not every line is another HeLa. The lessons are structural:

Immortal material outlives the consent behind it. HeLa cells have been used for over 70 years, across uses no one in 1951 imagined.
Genomic data creates new interests later. The 2013 dispute was about information that sequencing made available decades after collection.
Relatives have a stake. The people most affected in 2013 were descendants, who had never been asked anything.
Legality is not the end of the conversation. What was lawful in 1951 is now widely seen as wrong, and institutions have spent decades responding.

What did Moore v. Regents of the University of California decide?

Moore v. Regents of the University of California, decided by the Supreme Court of California in 1990, is the case most often cited on whether people own their removed cells. It is a state court decision, binding only in California, but it has been widely discussed far beyond it.

The facts, in general terms

John Moore was treated for hairy-cell leukemia at the UCLA Medical Center starting in 1976. His physician recommended removing his spleen, and over the following years took further samples during follow-up visits. According to the court’s opinion, the physician and a colleague planned research on Moore’s cells and had financial interests in that work, which they did not disclose to him. The University of California obtained a patent on a cell line derived from his cells in 1984, naming the researchers as inventors.

What the court held

The court reached two conclusions that are often summarized too simply:

Disclosure. Moore’s complaint stated a valid claim for breach of fiduciary duty or lack of informed consent. A physician must disclose personal research and economic interests that could affect medical judgment.
Property. Moore’s complaint did not state a valid claim for conversion, the legal wrong of interfering with someone’s property. The majority declined to extend conversion to excised cells used in research.

The majority gave policy reasons: patients were already protected by disclosure duties; conversion is strict liability and could reach innocent researchers far downstream; and uncertainty over ownership could slow research and disrupt the sharing of cell lines. It said the broader questions were better suited to the legislature. Justice Mosk dissented, arguing that Moore had a property interest and that the court could develop the law for new technologies.

Why it still matters

Moore is often read as “patients do not own their cells.” A more careful reading is narrower: in that case, the court rejected one property-based theory of liability, while protecting the patient through disclosure duties instead. The same tension runs through biocomputing. One side stresses that clear, transferable rights in cell lines make research and investment possible. The other asks why everyone in the chain can hold property in the material except the person it came from.

Did donors anticipate their cells being used for computing?

Consent for donated cells generally comes in three shapes.

Consent model	What the donor agrees to	Strengths	Weaknesses	Fit for biocomputing
Specific	A defined study or use	Donor knows exactly what will happen	New uses need fresh consent, often impossible	Clear, but rare for existing lines
Broad	A wide category of future research, often with oversight	Practical for biobanks and cell lines	Donors may not foresee uses like computing or commercialization	Common; ethically contested for novel uses
Dynamic	Ongoing, digital, revisable choices over time	Donors can opt in or out of new uses as they arise	Costly, needs reachable donors, risk of consent fatigue	Promising for new collections

Broad consent is the default for most cell banks. A 2015 workshop summarized by Grady and colleagues in the American Journal of Bioethics concluded that broad consent for future research can be ethically acceptable when paired with oversight and ongoing governance. In the United States, the revised Common Rule (45 CFR 46) added an optional broad consent pathway for storage and secondary research use of identifiable biospecimens and data.

For iPSCs specifically, Lowenthal and colleagues argued in Stem Cells Translational Medicine (2012) that donors should consent prospectively to a broad umbrella of foreseeable research, with recontact possible in limited circumstances. Boers and colleagues, writing in EMBO Reports (2016) about organoid biobanks, argued for pairing consent with ongoing governance, rather than treating a signature as the end of the donor’s involvement.

Dynamic consent, described by Kaye and colleagues in the European Journal of Human Genetics (2015) and discussed further by Budin-Ljøsne and colleagues (2017), uses a digital interface so participants can see how their samples are used and update their choices over time.

The foreseeability problem

The core question is whether “future research” can stretch to cover a commercial computing service. Reasonable people disagree.

The case that broad consent is enough: Donors agreed to open-ended research use. Biocomputing is research. Scientific uses are inherently unpredictable, and demanding foreseeability would freeze progress. Many donors give precisely because they want their cells to be useful in ways they cannot predict.

The case that it is not enough: Biocomputing differs from typical uses in ways a reasonable donor might care about. It involves brain tissue, which people regard as special. It may be commercial. It reframes the donor’s neurons as a component of a machine. The National Academies’ 2021 report on neural organoids noted that specific consent for organoid research was generally not obtained for past materials and that some donors might have objected had they been asked. Interviews with 60 patients and parents by Bollinger and colleagues (Stem Cell Reports, 2021) found broad support for organoid research, alongside a view that brain organoids are different from other organoids and conditions such as oversight and consent.

A middle position is common: broad consent may be lawful and acceptable for many uses, but for novel, sensitive, commercial uses, teams should go further where they can.

Can a “de-identified” cell line be traced back to its donor?

Every cell in an organoid carries the donor’s genome. Removing names from a cell line’s records does not remove the information written in its DNA.

Gymrek and colleagues showed in Science (2013) that the surnames of supposedly anonymous research participants could be inferred by matching markers on their Y chromosomes against public genealogy databases. Erlich and colleagues (Science, 2018) went further, showing that many people of European ancestry in the United States can be identified through relatives in consumer genetic databases, even if they have never taken a test themselves.

For biocomputing, this has three implications:

Sequencing creates new risk. A line used only for electrophysiology may never be sequenced. Quality control or research may later sequence it, creating identifiable data the donor never thought about.
Neural data is usually lower risk, but not zero. Electrical recordings from an organoid do not obviously reveal identity. Metadata linking recordings to a named line, plus a public genome for that line, could.
Relatives are exposed too. As the 2013 HeLa dispute showed, a genome speaks about family members who never consented to anything.

Open data and privacy can coexist. Standards like Neurodata Without Borders let labs share electrophysiology recordings with rich metadata about protocols and consent scope, while keeping genomic data under separate, controlled access.

Can a cell line, or a living computer, be patented?

Patent law varies by country, and this section describes the landscape only in general terms.

In the United States, the Supreme Court held in Association for Molecular Pathology v. Myriad Genetics (2013) that naturally occurring DNA sequences are not patent-eligible merely because they have been isolated, while synthetic complementary DNA (cDNA) can be. The general principle is that products of nature are not patentable, but human-made inventions involving biological material can be.

In practice, intellectual property in a living computer tends to attach to what people built, not to the donor’s biology as such:

Methods for reprogramming, differentiating and maturing cells
Specific engineered or modified cell lines
Electrode arrays, fluidics and other hardware
Control software, stimulation protocols and decoding algorithms
Trade secrets about culture conditions and quality control

Moore shows that a cell line derived from a patient can itself be the subject of a patent; the University of California held one on Moore’s. Whether that is fair is a separate question from whether it is lawful. Critics argue that patent systems reward the final transformation and ignore the original contribution. Defenders reply that the donor contributed raw material available in every human body, while the inventors contributed the scarce skill that made it valuable.

Benefit sharing does not have to mean cash. Options discussed in bioethics include:

Transparency: telling donors and the public how lines are used and by whom.
Governance roles: giving donors or their representatives a voice, as the NIH-Lacks working group did for HeLa genome data.
Community benefits: funding health programs, education or research priorities chosen with donor communities.
Open science: publishing data and methods so the public benefits from what donated cells made possible.
Recognition: acknowledging donors, where they wish, rather than erasing them.

Community engagement matters especially where groups have historical reasons to distrust research. The National Academies’ 2021 report noted that past research abuses involving African American and Native American participants make consent questions particularly sensitive.

The case for donor compensation or royalties

Fairness: if a line generates revenue, it seems unjust that everyone profits except the source.
Respect: payment acknowledges that donors contributed something of value, not just waste tissue.
Trust: visible sharing could increase willingness to donate, especially in underrepresented communities.
Consistency: people are paid for plasma and research participation in many places, so the principle is not foreign to medicine.

The case against

Undue inducement: payment, especially royalties with uncertain upside, could pressure financially vulnerable people to donate.
Attribution is hard: one line is rarely the decisive input. A computing service may use many lines, and most value comes from engineering.
Friction for research: royalty obligations attached to lines could make sharing slower and more expensive, the concern the Moore majority raised.
Commodification: treating body parts as revenue-generating property may itself erode the gift ethic that much tissue donation rests on.

A frequent compromise is clear disclosure plus community-level sharing: tell donors honestly that commercial use may occur and that they will not receive royalties (if that is the case), and direct some benefit to donor communities or public goods. Others argue this is a polite way of keeping the status quo. The debate is not settled, and it should not be presented as if it were.

Withdrawal is where cell lines differ most from ordinary research participation. A survey participant can stop answering questions. A cell line keeps dividing after the donor has left.

In practice, withdrawal from cell line research usually means:

Stopping future distribution of the line by the repository.
Destroying stored samples under the repository’s control.
Removing data from future releases where feasible.

It usually cannot mean recalling cells already shipped to other institutions, deleting data already shared, or retracting published results. Bioethicists have long debated what a right to withdraw means for biobank research, for example Eriksson and Helgesson in the European Journal of Human Genetics (2005).

For biocomputing, withdrawal adds a new wrinkle. If a company has built models, benchmarks or datasets from a particular line, what happens to them? Reasonable answers differ. A strict view says derived outputs should be retired where practical. A pragmatic view says outputs are new works and only future use of the cells must stop. Either way, the honest approach is to tell donors before they agree.

What should responsible biocomputing companies do?

No company can retroactively make every existing line ideal. But some practices are within reach today. These are recommendations, not legal requirements:

Track provenance for every line. Record the source, consent form version, permitted uses, commercial terms, withdrawal procedures and any restrictions, and attach that record to every experiment.
Prefer lines with explicit consent covering neural differentiation, data sharing and commercial use. Seek ethics review before using lines whose consent is ambiguous.
Make consent machine-readable. Encode consent scope as metadata that travels with datasets, so downstream users inherit restrictions automatically.
Re-consent where feasible. When donors can be reached through the original institution, offer a clear explanation of biocomputing and a fresh choice. For new collections, consider dynamic consent.
Protect genomic data. Keep sequences under controlled access, separate from openly shared neural recordings, and avoid publishing raw genomes without consent.
Publish transparency reports. Say which lines are in use, under what consent categories, how many withdrawal requests were received and how they were handled, without identifying donors.
Engage donor communities. Create advisory roles for donors or patient groups, especially for lines from historically marginalized populations.
Be explicit about benefits. If donors will not receive royalties, say so plainly at consent. If the company shares benefits at community level, describe how.

These steps line up with broader calls in the field. The Baltimore declaration (Hartung et al., 2023) and the organoid intelligence roadmap by Smirnova and colleagues (2023) both call for ethics to be built into the research program from the start, and the ISSCR guidelines set expectations for informed consent when donating materials.

How do the main positions compare?

Most debates about owning living computers reduce to a few positions. Each has serious defenders.

The property view: donors should hold real property rights in their cells, including a share of commercial value. Strength: fairness and control. Weakness: legal complexity and possible chilling effects on research.
The consent-and-disclosure view: donors hold rights to be informed and to refuse, but not ownership; this is roughly where Moore landed. Strength: workable and established. Weakness: may leave donors with little once consent is signed.
The stewardship view: institutions hold cell lines as trustees with ongoing duties to donors and communities. Strength: matches long-lived, shared resources. Weakness: depends on institutions honoring duties no one enforces.
The public-good view: donated cells should mainly support open, non-exclusive science. Strength: maximizes shared benefit. Weakness: may not attract the investment needed to build products.

Most companies will operate in a blend: legally under consent and contract, ethically under some form of stewardship.

What questions remain open?

Does broad consent cover computing? Ethics bodies have not reached consensus on whether specific consent should be required for brain organoid or computational uses.
Who speaks for donors who cannot be reached? For older lines, recontact is often impossible. Community representation is one answer, but whose community?
How should value be attributed? If a service uses many lines, is there a fair way to share benefits among them?
What happens to derived outputs after withdrawal? There is no shared standard for models or datasets built from a line whose donor later withdraws.
Will law change? Courts and legislatures could revisit ownership of biological materials as products built on them grow more valuable.
Do relatives have rights? Genomic data implicates family members, yet most consent frameworks address only the donor.
How will these questions interact with moral status? If future organoids were ever judged to have interests of their own, as explored in the moral status of biological computers, “ownership” might need rethinking entirely. Our piece on the future of biocomputing looks at where the technology may go.

So, who owns a living computer?

Who owns a living computer is less a single answer than a set of obligations. Under current US law, shaped by Moore, donors generally do not hold property rights in their excised cells, while companies own what they build and customers own their data under contract. The HeLa story shows why that legal answer is not the whole ethical one. Immortal cells outlive consent, genomes speak for relatives, and trust depends on whether donors can see what was done with their gift.

The practical path forward is unglamorous: provenance records, clear consent, protected genomes, honest benefit terms and public reporting. At Organoid Programming, we are building provenance and consent metadata into the software layer of organoid programming itself, so that the origin of every culture travels with every experiment. You can read more about our approach on our ethics page.

Frequently asked questions

Who owns a biocomputer made from donated human cells?

Usually no single party owns everything. In the United States, the leading case, Moore v. Regents of the University of California (1990), rejected a patient’s claim that he kept a property interest in his excised cells for the purposes of a conversion lawsuit. In practice, ownership is set by contracts: repositories license cell lines, companies own the hardware, software and their own inventions, and customers usually own their data. Donors typically hold consent-based interests rather than property rights.

Did Henrietta Lacks consent to the use of her cells?

No. In 1951, while Henrietta Lacks was being treated for cervical cancer at Johns Hopkins, doctors took tumor cells without her knowledge or permission. Those cells became the HeLa line. Taking them was legal at the time, and Johns Hopkins says it would not happen today without consent and that it never profited from the cells. Her family learned about the line years later.

What was the 2013 NIH agreement with the Lacks family?

In 2013, after a HeLa genome was published without the family’s involvement, the US National Institutes of Health reached an agreement with the Lacks family, reported by Nature. Full HeLa genome data would be held under controlled access, and researchers would apply to use it, with requests reviewed by a working group that includes members of the Lacks family. It gave the family a voice, but no ownership or payment.

Can a cell donor withdraw consent after a cell line is made?

Usually only partly. Many consent forms let donors stop future use, so a repository can halt new distribution of a line and destroy stored samples. But cells already shipped to other labs, data already shared and results already published generally cannot be recalled. Responsible consent forms explain these limits up front, so donors know what withdrawal can and cannot do before they agree.

Should cell donors get royalties from biocomputing?

There is no consensus. Supporters argue royalties recognize the donor’s contribution and correct an imbalance where everyone profits except the source. Critics worry that payment could pressure people to donate, complicate research sharing, and value a single donor’s cells far above what any one line contributes. Many ethicists favor transparency and community-level benefit sharing over individual royalties, though that view is itself debated.

Can anonymous cell lines be traced back to donors?

Sometimes, yes. Every cell carries the donor’s genome. Gymrek et al. showed in Science (2013) that surnames of anonymous research participants could be inferred by matching Y-chromosome markers to public genealogy databases, and Erlich et al. (2018) showed that many Americans of European ancestry can be identified through relatives in consumer genetic databases. Biocomputing companies should treat genomic data from cell lines as potentially identifiable.

Did donors of existing iPSC lines agree to biocomputing?

Mostly not explicitly. Many induced pluripotent stem cell lines were collected under broad research consent before brain organoids or biocomputing were widely discussed. Broad consent may legally cover new uses, but whether donors would have agreed to their neurons running computations for a commercial service is a separate ethical question. Responsible teams check each line’s consent scope and prefer lines with clear, recent language.

What should a responsible biocomputing company do about donor consent?

Recommended practices include tracking provenance for every line (source, consent version, permitted uses, commercial terms), preferring lines whose consent covers neural and commercial use, offering re-consent or dynamic consent where donors can be reached, protecting genomic data, publishing regular transparency reports, and engaging donor communities. None of these settles ownership, but together they make claims checkable and keep donors’ interests visible.

Who owns a living computer? Donor consent, cell lines and the lessons of HeLa

What is a “living computer,” and why is ownership a hard question?

Who claims what, from donor to computing service?