There are some jobs where replacing a human with a robot feels fairly unremarkable. Building a car? Fine. Shifting boxes around a warehouse? We’ve seen it before. Injecting a single sperm cell into a human egg to help create an embryo? That one’s a little more novel, and it’s already happening, thanks to the advent of AI IVF robots.
One of the most ambitious attempts comes from New York-based Conceivable Life Sciences, which has spent the past few years developing AURA – an automation-assisted IVF laboratory designed to carry out much of the painstaking work traditionally performed by highly trained embryologists. It combines robotics, computer vision, algorithms, and AI to prepare sperm, identify and handle eggs, and perform the microscopic injection involved in fertilisation.
Inconceivable? Far from it. Bloomberg reported in January that Conceivable’s technology had already helped bring 19 babies into the world. More recent peer-reviewed research also gives us a better idea of what some of Conceivable’s earlier robotic systems can actually do.
Before you start picturing robot nurseries, though, there’s an important distinction. AURA isn’t an artificial womb, AI isn’t designing babies, and robots aren’t carrying pregnancies. At least, not any time soon. Instead, Conceivable is trying to automate a particularly delicate part of IVF that already takes place in labs.
But the really interesting question isn’t whether a robot can do it at all. It’s whether it could eventually do it better than a human…
Making an embryo is incredibly fiddly work
In vitro fertilisation or IVF might be a familiar acronym, but what actually happens in an embryology lab may not be. Following ovarian stimulation and egg retrieval, embryologists have to identify and prepare eggs and sperm before attempting fertilisation. One commonly used technique is intracytoplasmic sperm injection, better known as ICSI, where a single sperm is injected directly into an egg.
A conventional procedure can involve an embryologist examining sperm under a microscope, selecting one, and immobilising it before drawing it into a tiny pipette. Eggs retrieved in follicular fluid also have to be found, isolated, and prepared. During ICSI, the embryologist holds an egg in place, carefully penetrates its outer membrane, and injects the sperm. The fertilised egg is then cultured to see whether it develops successfully, eventually reaching the blastocyst stage, when an embryo may be suitable for transfer or freezing.
If it wasn’t obvious by now, it’s highly skilled, super-fiddly work, which is precisely why Conceivable thinks there’s room for automation. Its current AURA platform consists of interconnected automated workstations designed to cover processes from preparing dishes and sperm to handling eggs, ICSI, embryo culture, and vitrification – the rapid-freezing process used to preserve eggs and embryos.
Computer vision can also identify eggs among follicular fluid, while automated systems prepare sperm and robotic micromanipulation tools handle procedures that are normally carried out by an embryologist. Conceivable says that AURA covers more than 200 individual steps involved in taking egg and sperm through to an embryo.
Importantly, the company also describes AURA as automation-assisted, rather than an autonomous replacement for an embryologist – clinical decisions remain with humans.
Where does AI come into IVF?
It’s tempting to slap an “AI robot” label onto the AURA and call it a day, but that would make things sound simpler than they really are. More accurately, it’s better described as a collection of different tech working together. Precision robotics provide the physical movements, computer vision identifies biological material, and algorithms can help with jobs such as finding and selecting sperm and eggs.
Different parts of Conceivable’s experimental systems have also operated with varying levels of autonomy. In a peer-reviewed study published in Human Reproduction, researchers tested earlier Conceivable prototype systems nicknamed Pearls. One prepared sperm, another identified and isolated eggs and removed their surrounding cells, while a third selected sperm, immobilised it with a laser, and performed ICSI.
Multiple AI systems were used across those machines, but some operations still needed direct intervention or digital control from a human. Full task autonomy was achieved only for sperm preparation and selected parts of ICSI.
In other words, this isn’t a fertility clinic pressing a big green button and coming back nine months later to collect a baby. What Conceivable is aiming for seems more like an extraordinarily sophisticated robotic production line, with human specialists still overseeing what’s going on.
Why use robots to make human embryos at all?
For anyone going through IVF, the idea of handing even more of the process over to machines could, understandably, feel uncomfortable. These aren’t just cells moving through a production line, after all, and behind every procedure is someone hoping to start or grow a family.
The argument for automation isn’t that an algorithm understands reproduction better than an embryologist, or that removing humans from the process is somehow desirable. It’s much more practical than that. Some of the lab work involved in IVF demands extraordinary precision and consistency, and robots could potentially perform the same microscopic movements in exactly the same way, again and again. We’ve seen robots outperform our clunky meat-covered skeletons countless times, in everything from sports and manufacturing, to demos at tradeshows, and the same precision improvements apply here.
Current IVF laboratories are still heavily dependent on manual procedures. Technique can vary between people and laboratories, while workload, experience, and fatigue can contribute even more. In fact, a 2026 review of IVF automation published in the Journal of Assisted Reproduction and Genetics describes operator-dependent procedures, including embryo selection, vitrification, and ICSI, as major targets for automation.
Apart from accuracy, machines can also record huge amounts of information about what they’re doing. An experienced embryologist knows how to manipulate an egg, but they aren’t recording the precise movement, position, timing, and other microscopic variables involved every time they do it. A robotic system potentially can. Combine that information with the outcomes of thousands of attempts, and, in theory, you can start finding tiny differences that correlate with better results and refine the process accordingly.
The bigger prize, though, could simply be scale. The World Health Organization estimates that around one in six people of reproductive age worldwide experience infertility during their lifetime, while access to treatment remains severely limited. Its first global infertility guideline, published in late 2025, highlighted cost as a major barrier, with a single round of IVF costing as much as twice the average annual household income in some settings.
Conceivable’s long-term plan could see highly automated IVF superlabs, where relatively small numbers of specialists could supervise far greater numbers of procedures. In theory, that could mean making IVF available to more people, which is another massive win.
Are AI IVF robots actually better than humans?
For all the promise of greater precision and consistency, there’s still not enough evidence that Conceivable’s technology can outperform human embryologists.
Its peer-reviewed proof-of-concept study involved just 11 cases in which two or three of its earlier automated systems were used sequentially. The automated procedures achieved fertilisation in 64.3 per cent of injected eggs, compared with 81 per cent for the manual procedures. Of the resulting fertilised eggs, 42.2 per cent developed into usable blastocysts in the automated group, compared with 59.6 per cent in the manual group.
Overall, twelve single-embryo transfers from the automated arm ultimately resulted in five healthy live births. That’s still an impressive milestone, and it shows that several extremely delicate IVF procedures can be carried out sequentially using automated systems and result in healthy babies. What it doesn’t show, though, is that robots are better at IVF than humans.
It’s worth noting that the researchers themselves caution that the study was too small for a statistical comparison between the automated and manual approaches. It was also sponsored by Conceivable, many of its authors work for or hold financial interests in the company, and some procedures still required human intervention.
That June 2026 review mentioned earlier, by the way, also puts these results into perspective. The study was small and funded by Conceivable, and the eggs handled manually by embryologists actually produced better results than those handled by the automated systems – although there weren’t enough participants to draw firm conclusions from that comparison.
It’s much bigger than one company
AURA might be one of the more dramatic examples, but AI and automation are already spreading through reproductive medicine.
Algorithms can already analyse images of developing embryos and try to rank which are the most promising. Time-lapse incubators can also continuously photograph embryos as they develop rather than requiring them to be periodically removed for assessment. Researchers are also investigating automated sperm analysis and selection, vitrification, warming, and ICSI.
A major 2026 review in Nature Reviews Bioengineering also argues that automation, computer vision, and deep learning could eventually come together in something approaching an AI-integrated, closed-loop IVF laboratory. But there’s still a big difference between promising tech, and actually helping more people have babies.
The UK’s Human Fertilisation and Embryology Authority (HFEA), for example, currently gives time-lapse incubation using automated embryo analysis – including algorithms or AI – a black rating for improving the chances of having a baby for most fertility patients. In the HFEA’s system, that means moderate-to-high-quality evidence shows the treatment add-on has no effect on that outcome.
Ultimately, being better at analysing an image doesn’t automatically translate into more pregnancies or live births, and that’s absolutely worth remembering whenever AI and IVF appear in the same sentence – especially if you’re understandably extremely emotionally invested in the outcome.
Could AI robots make IVF cheaper?
In theory, reducing the amount of highly specialised manual labour required for every procedure could allow labs to process more cases. Standardising procedures might reduce mistakes, while better outcomes could mean that some patients need fewer expensive treatment cycles.
But even if automated labs save money, there’s no guarantee those savings will be passed on to patients. The price of IVF isn’t simply the cost of moving a pipette, after all. Medication, consultations, egg retrieval, embryo transfer, testing, storage, clinic overheads, and plenty of other costs contribute to the final bill.
Even a dramatically more efficient laboratory doesn’t automatically mean dramatically cheaper treatment. And remember – everything that led to the creation of this technology has had millions pumped into it. And investors always want their money back. With interest.
The most likely future for this technology, then, isn’t an AI robot replacing every single step in the process. Instead, embryologists could increasingly become supervisors and problem-solvers, stepping in when biology doesn’t behave the way a machine expects.
For now, though, the humans aren’t going anywhere. While robots have already proved that they can help produce embryos that become healthy babies, proving that they can do it better, more cheaply, and at the scale that Conceivable envisages, is the much bigger challenge.
Read the full article here
