Recently it was announced the first artificial wombs facility. The press release included a link to a video explaining how this disruptive company would operate in terms of genetic engineering, biotechnology, and an artificial intelligence (AI) system for artificial wombs, promising solutions to issues related to human reproduction.
Artificial wombs are not a recent study field in scientific research. In the last few years, researchers were able to grow fetal lambs placed in an extra-uterine device filled with synthetic amniotic fluid for up to 4 weeks and develop a system for robustly growing normal mouse embryos ex-utero. In addition, a European university has received a large EU grant to build an artificial womb adequate to substitute the protective environment of the maternal womb in case of premature birth. Despite the scientific research, progress and investment in the area, an artificial womb capable of growing a human embryo from the very beginning still needs to be developed. The video included in the abovementioned press release merely shows a concept design by Hashem Al-Ghaili based on innovative research conducted by scientists and engineers since 1950. Hashem explains that the technology to create artificial wombs already exists, and that his envisioned medical complex could be ready in 10 to 15 years. He claims that research restrictions on human embryosbeyond 14 days is what has held back innovation in this field. Some scientists in humanreproduction agree that the technology presented is not so absurd and could become a reality.
Nevertheless, as regards the European Union legal framework, it is not only the prohibition on human embryo research that should be considered. Other relevant legislation must also be upheld.
According to the video,every artificial womb pod would have a system controlled by artificial intelligence to monitor vital signs and developmental progress in real-time, guaranteeing that each baby receives custom nutrients tailored to their needs, replicating the exact conditions inside the mother’s uterus. This AI system will also be able to detect, inform, and provide medical guidance on the baby’s physical features and potential genetic abnormalities. Parents would receive real-time monitoring reports on a mobile app and the technology would allow them to interact with their babies and follow their development. The artificial womb, as described above, is a very complex medical device that integrates different types of technology, including AI. Considering that it could become a reality soon, certain regulatory demands should be observed. Therefore, besides the prohibition of research on human embryos over 14 days – which already raises intense ethical and moral debates – some aspects regarding the AI used in the medical device system must be analysed under the existing and future EU regulatory framework.
The Charter of Fundamental Rights of the European Union states in its preamble the importance of protecting fundamental and human rights from changes in society, social progress, and scientific and technological developments. In the European Union, regulation is built upon a strong respect for fundamental and human rights, which requires compliance in technological and healthcare innovation.
According to Hashem Al-Ghaili’s concept, and as described above, an AI system would monitor babies’ growth, provide medical guidance and reports for clinicians, and allow parents to participate in their baby’s development process. Therefore, the artificial womb pods could be defined according to Article 2 (1) of the Regulations on Medical Devices (Regulation (EU) 2017/745) and Article 2 (2) of the In Vitro Diagnostic Devices Regulation (EU) 2017/746, having to comply with the requirements under both sets of legislation. Another regulation that would be applicable to this technology is the Artificial Intelligence Act, once it comes into force. The European Commission’s Proposal for the Artificial Intelligence Act puts forth specific objectives to guarantee that artificial intelligence systems placed on the market are aligned with the Union’s legislation and values, ensure legal certainty to encourage investment and innovation, enhance governance and the effective enforcement of existing legislation on fundamental rights and security requirements, facilitate the development of a single market for legal, safe and reliable applications of AI and avoid market fragmentation. This proposal is part of the EU’s broader strategy to promote the development and use of trustworthy AI and to ensure that it is used with respect for fundamental rights and values, including privacy and data protection, non-discrimination, and transparency. The AI Act classifies AI medical devices according to the potential damage they can cause to patients. Under the AI Act, the artificial womb’s artificial intelligence system could be classified as high risk when fulfilling both of the following conditions: (i) the AI system is intended to be used as a safety component of a product or is itself a product; (ii) a product whose safety component is the AI system or which is itself the AI system. These systems are required to undergo a third-party conformity assessment with a view to their placing on the market or putting into service. The
requirements include compliance; risk management system; data and data governance; technical documentation; record keeping; transparency and provision of information to users; human oversight; accuracy, robustness and cybersecurity. Although there is no specific provision in the AI Act proposal for the use of artificial intelligence in healthcare, in June 2022 the European Parliament published a multidisciplinary study carried out by PwC to evaluate potential uses: “Artificial Intelligence in Healthcare”. This research provided the primary basis for our analysis of artificial womb technology and identification of the risks involved, which will need to be mitigated before this AI system can be put into service, if it ever becomes a reality.
One of the challenges faced relates to the database needed for the AI system to function. Because the applicable legislation includes restrictions on embryo research, it would not be easy to provide a database on babies’ development in artificial wombs for the AI system to function correctly. Given that every growth pod would have to rely on AI to replicate the exact conditions of a natural womb and to provide medical guidance and reports on babies’ developmental progress, an inadequate database would represent potential harm to babies’ development due to AI errors and would increase the risk of bias and perpetuation of inequalities. Furthermore, the risk of data breaches and cyberattacks that could expose sensitive information would need to be correctly addressed, particularly since these babies would be subject to personal data collection from the very beginning of their lives. Consequently, the first significant step that needs to be overcome for the AI system of the artificial womb to work correctly and become a reality is making a sufficiently reliable, robust, secure, trustworthy and traceable database so that the AI system can work correctly without posing risks to the babies.
The Artificial Intelligence in Healthcare report suggests some mitigation measures for these problems, such as the definition of standardised methods and procedures for the extensive evaluation and regulatory approval of AI solutions, particularly regarding their generalisations to new populations and sensitivity to potential errors on database inputs. Also, the AI algorithms should be designed and implemented as assistive tools (as opposed to fully autonomous tools) so that clinicians remain part of the data processing workflow to detect and report any errors or contextual changes and to minimise potential harm to patients. To reduce bias, AI training should be provided using balanced, representative datasets and interdisciplinary approaches to medical AI involving social scientists, with a view to promoting greater diversity and inclusion in medical AI considering the existing genetic differences in the world’s population. In addition, it is of utmost importance to secure data privacy and make significant investments in cybersecurity. Nevertheless, the prohibition of research on embryos after 14 days would have to be removed. From a scientific standpoint, a technological innovation capable of carrying out the complete development of a human being, providing real-time medical care, and allowing parents to feel closer to their baby through an app, among other solutions, is a cutting-edge device. However, from an ethical and legal standpoint, several issues must be considered before artificial wombs can be effectively implemented. Although it is not yet possible to develop a human being from scratch due to legal restrictions, the artificial intelligence system described in this Insight has the potential to revolutionise not only human reproduction, but also the medical treatment of premature babies and neonatal care. This AI system would allow parents to connect with their babies under medical supervision. It would also act as an assistant tool, helping medical teams provide better health diagnoses and solutions in this field. Although several legal assessments are necessary before this artificial intelligence system can be used for these means, if all due diligence is done to mitigate risks, this technology could transform healthcare. Technological development, especially in healthcare, has the potential to support progress and provide intelligent solutions to global health issues. However, its exponential growth without careful assessment may expose society to unnecessary risks, disparities, and inequalities. Regulation’s primary objective must be to find a fair balance between protecting life in all its forms and encouraging technological innovation to ensure humanity’s prosperous and just future.
