Note: The ethical judgments on this page refer exclusively to the action — never to the person who performs it or who came into existence through it. Cf. Note on Ethical Judgments.
An automated reproduction laboratory is an artificial agent that carries out the working steps of an assisted reproduction laboratory — gamete preparation, fertilization, culture, assessment, cryopreservation — wholly or partly without human performance. What is meant is not a single device but the chaining of the steps into a procedure that runs from the oocyte to the stored embryo without a hand having to intervene.
The ontological point lies in the fact that this system is value-neutral. It carries out what is assigned to it. A robotic arm that changes a culture medium differs morally in nothing from an embryologist who does the same. The machine becomes person-relevant only at a single point: where the assigned step is a decision about the existence of a person.
Ontological classification
- is a subclass of: artificial agent
- presupposes: in vitro fertilization — and inherits its verdict along the path of presupposition
- carries out: algorithmic embryo assessment, insofar as the assessment is assigned to the system
- carries out: algorithmic embryo selection, insofar as the choice too is assigned to it
- stands in relation to: nominal oversight as the shape that human control typically takes in automated operation
The laboratory as such is not treated here as an intrinsically evil act, and this is a reasoned decision. An action is determined by its object: by what the agent knowingly and willingly does. The object of automation is the transfer of a manual operation from a human being to a machine. In this, taken by itself, there lies no disregard of a person. The verdict therefore does not sit on the technique but on the individual steps that are assigned to it.
What is actually automated
The sober state of affairs is to be separated from the narrative that accompanies it.
There are announced fully automated systems; the first, AURA, was announced in July 2026 for clinical use from 2027. The published degree of autonomy lay in this case between about sixty and seventy percent per working step — oocyte search, washing, denudation, injection — and the rest ran under human remote control. The authors of the corresponding study themselves recorded that full autonomy had been achieved only for sperm preparation and for individual sub-steps of the injection.
For the ethical debate one detail is decisive: the embryo assessment in the published study was still manual. Precisely the step at which the criticism is kindled was not in use at all. Whoever speaks of automated reproduction laboratories therefore speaks so far of two different things — of a partly automated handling that exists, and of an algorithmically performed choice that has been announced.
To this is added a finding that disturbs the obvious narrative of progress: the only controlled comparative series came out to the disadvantage of automation. On the same cohorts of oocytes, the fertilization rate in the automated arm lay clearly below that of the manual one; the number of usable blastocysts likewise; and the automated procedure took a multiple of the time. The embryologists beat the machine at both intermediate stages.
From this no moral judgment follows, but a rule of caution does: what the available evidence supports is standardization and error reduction; what is regularly promised, by contrast, is better clinical outcomes. These are two different promises, and only the first is covered.
Where the machine becomes person-relevant
The ontology of personhood examines not whether a procedure is novel or unnatural, but at which points it touches persons and how it treats them. Ontological dignity is the inalienable, objective worth of the person and the sufficient ground of the personalistic norm: the person is to be affirmed for her own sake. This norm is violated by instrumentalization and by forgetfulness of the person — the affirmation of a person not for her own sake but on account of her characteristics.
If one applies this standard to the sequence of steps of an automated laboratory, a clear separation results.
The handling steps. Pipetting, washing, changing the medium, freezing, transporting: here the system touches an embryo as a person in the First Dimension, but it decides nothing about him. What must be examined in the machine is diligence; the question of moral admissibility does not arise anew here, because it is already decided at the presupposed procedure — and decided for the human embryologist no less.
The assessing step. An algorithmic embryo assessment is a temporal attribution and hence precisely not essence-determining. It says something about image features and a statistical probability of implantation, nothing about the status of the one it assesses.
The choosing step. Here and only here the matter tips. If a decision about which embryo is transferred and which is not is based on such an attribution, then a temporally limited, non-essence-determining assessment becomes the ground of a decision about existence. Algorithmic embryo selection is therefore at once eugenic selection and an automated decision process. And here lies the decisive finding: automation adds no new moral quality to selection — it withdraws the personal performance from it. What is added is the responsibility gap, not the selection.
What holds independently of the technology
One finding applies to the automated laboratory irrespective of which steps it takes over: it constructively presupposes in vitro fertilization. In the ontology of personhood this is carried, by way of artificial fertilization, as an intrinsically evil act and as a case of practical forgetfulness of the person: generation is detached from the personal act of the spouses and handed over to the disposal of third parties, and the procedure deliberately brings forth more embryos than are transferred.
The evaluation therefore travels not by inheritance but along the path of presupposition — and it applies even where no algorithm selects and every operation is carried out with care. To examine the automated laboratory solely by how reliably it works and who supervises it would be to begin the examination halfway through: the procedure whose steps are being automated stands objected to before any automation. In magisterial terms this is the statement of Donum vitae II, B, 5 concerning the so-called simple case — a homologous fertilization free of any destruction of embryos remains illicit, because it deprives generation of the dignity proper to it.
The value-neutrality of the system is thereby precisely determined and at the same time limited: it concerns the technology, not the practice into which the technology is placed.
This has an uncomfortable reverse side for both camps. Whoever condemns the machine but holds the choice made by hand to be unproblematic has missed the object. And whoever wants to cure automation by better oversight cures the gap, not the discarding.
The balance of errors
One argument deserves attention of its own, because it cuts in both directions. Automation trades distributed, frequent, small human errors for rare, correlated, potentially catastrophic system errors. Both kinds of error are documented: mix-ups and handling errors on the one side, technical failure with the loss of entire stored holdings on the other.
The difference is not the frequency but the correlation. Human errors are independent of one another; a system error strikes everything subject to the same procedure at the same time. Whoever draws up the balance must therefore not compare error rates but distributions of magnitudes of damage — and the balance is not obvious in either direction. It is neither obviously positive nor obviously negative, and every account that presents it as obvious has left out the harder part of the calculation.
What can be regulated and what cannot
The objections to an automated reproduction laboratory fall into two classes, and the distinction is not classificatory but practical: it decides whether an objection can be answered by a better procedure.
Consequence-based objections are directed against effects and attendant circumstances: biased training data, lacking explainability, diffusion of responsibility, inadequate information about which steps are automated. They lapse with their cause and are therefore curable by a governance framework.
Act-based objections are directed against the action itself, independently of diligence and circumstances: the discarding by characteristics and, within the magisterial tradition, the detachment of procreation from the personal act of the spouses. These objections obtain even when the system works faultlessly, transparently and under seamless oversight.
From this follows the systematic thesis: what a governance framework answers is without exception a consequence-based objection. Where an act-based objection obtains and only a body of rules is offered, the question of the admissibility of the practice is not answered but passed over.
The strongest objection to this assessment
The most serious contradiction comes not from enthusiasm for technology but from empirical evidence: standardization is a genuine good where the spread between good and bad laboratories is large.
And it is large. Even human embryo assessment according to the customary reference scheme achieves only moderate agreement between two experienced assessors; between centres of differing quality the gap is greater than any difference a machine has ever made. Whoever measures automation against top laboratories measures it against the best case and misses the comparison that matters: not machine against the most experienced embryologist, but machine against the median of world practice. A procedure that is equally mediocre everywhere can in sum bring more children to birth than one that is excellent in a few places and bad in many.
This objection strikes — and it strikes exactly half the matter. It answers the consequence-based objections completely and the act-based ones not at all. A discarding by characteristics does not become something else by taking place more reliably, more reproducibly and in the same way in more clinics. On the contrary: precisely the reliability that the standardization argument promises turns a local practice into a comprehensive order. The aggregation argument begins here — not that anyone would be coerced, but that the sum of free and in each case well-justified decisions produces an order in which certain modes of existence count as an avoidable error.
A second point belongs here, because it attacks the argument’s own footing: the reference to the weak evidence has an expiry date. “It does not even work” is a strong argument today and possibly obsolete in ten years. Whoever bases his critique on this alone loses his position as soon as predictive power increases. The normative argument must carry independently of this — and it does, because it does not hang on the hit rate but on what a hit rate cannot say about a person.
See also
- Algorithmic Embryo Selection — the step at which it tips
- Algorithmic Embryo Assessment
- Nominal Oversight, Governance Framework
- In Vitro Fertilization, Embryo Transfer
- Preimplantation Genetic Diagnosis, Eugenic Selection
- Supernumerary Embryo, Embryo
- Responsibility Gap, Degree of Autonomy
- AI-Algorithmic Arrangement, Forgetfulness of the Person
Sources: Generated by querying the ontology of personhood. Research as of 29 July 2026.
Further sources:
- Chávez-Badiola, A., Cohen, J. et al. (2026): Automated oocyte retrieval, denudation, sperm preparation, and ICSI in the IVF laboratory: a proof-of-concept study and report of the first live births. Human Reproduction 41(2): 214–230.
- Illingworth, P. J., Venetis, C., Gardner, D. K., Nelson, S. M. et al. (2024): Deep learning versus manual morphology-based embryo selection in IVF: a randomized, double-blind noninferiority trial. Nature Medicine 30(11): 3114–3120.
- Zaninovic, N., Sierra, J. T., Malmsten, J. E., Rosenwaks, Z. (2024): Embryo ranking agreement between embryologists and artificial intelligence algorithms. F&S Science 5(1): 50–57.
- Thirumalaraju, P., Kanakasabapathy, M. K. et al. (2026): Stability and reliability of artificial intelligence models in embryo selection for in vitro fertilization. Fertility and Sterility 125(2): 277–286.
- Koplin, J. J., Johnston, M., Webb, A. N. S., Whittaker, A., Mills, C. (2025): Ethics of artificial intelligence in embryo assessment: mapping the terrain. Human Reproduction 40(2): 179–185.
- Human Fertilisation and Embryology Authority: Rating of IVF add-ons (traffic-light scale) as well as documents of the scientific advisory committee on automation and robotics in the laboratory.
- Congregation for the Doctrine of the Faith (1987): Donum vitae, II,B,4–5.
- Congregation for the Doctrine of the Faith (2008): Dignitas personae, nn. 14, 17, 22.
- Dicasteries for the Doctrine of the Faith and for Culture and Education (2025): Antiqua et nova, n. 74.