Advances and Current Practices in Assisted Reproductive Technology: Integrating Artificial Intelligence

Infertility affects an estimated 10–15% of couples worldwide, positioning assisted reproductive technology (ART) as a cornerstone of contemporary fertility care [1,10]. Since the landmark birth of the first in vitro fertilization (IVF) baby in 1978, more than 10 million children have been conceived through ART, underscoring its profound global impact [11]. Conventional ART techniques—such as IVF and intracytoplasmic sperm injection (ICSI)—have evolved substantially with the integration of time-lapse embryo imaging, preimplantation genetic testing (PGT), and advanced cryopreservation methods. These innovations have collectively improved implantation and live birth rates while markedly reducing the incidence of multiple gestations [6]. Despite these advances, ART-conceived pregnancies continue to exhibit higher risks of hypertensive disorders, pre-eclampsia, preterm delivery, and low birth weight, even in singleton births [12,13]. These adverse outcomes are likely multifactorial, reflecting the interplay between maternal characteristics, hormonal influences, and altered placentation associated with assisted conception cycles [14]. Long-term follow-up studies of ART-conceived offspring are largely reassuring, indicating normal neurodevelopmental outcomes overall. However, subtle cardiometabolic variations—such as elevated blood pressure and endothelial alterations—have been observed during adolescence, suggesting potential epigenetic or developmental influences [15,16]. Concurrently, the field of ART is being transformed by emerging technologies, including artificial intelligence (AI)-driven gamete and embryo selection, mitochondrial replacement therapy (MRT), and in vitro gametogenesis (IVG). The integration of machine learning algorithms into embryo assessment and individualized ovarian stimulation protocols, alongside advances in vitrification systems, is shifting ART toward a more data-informed and patient-tailored discipline [17,18]. These innovations hold promise for enhancing embryo viability prediction and minimizing complications such as ovarian hyperstimulation, though their widespread implementation requires robust clinical validation and long-term safety evaluation [19]. In light of these developments, this review critically examines recent evidence from 2015–2025 on ART outcomes. It focuses on three key domains: (1) maternal and perinatal outcomes and their underlying biological mechanisms; (2) long-term health and developmental trajectories of ART-conceived children; and (3) optimization of ART protocols through personalized medicine and advanced technologies. Synthesizing current findings across these areas will clarify ART’s evolving benefits, limitations, and the ethical and clinical imperatives guiding its future application.

METHODS

This review was conducted using a narrative review methodology. Recent literature published between 2015 and 2025 was searched across PubMed, Scopus, and Google Scholar using keywords such as “assisted reproductive technology,” “AI in ART,” “mitochondrial replacement therapy,” “in vitro gametogenesis,” and “artificial ovaries.” Peer-reviewed articles, clinical trial reports, and guidelines from authoritative regulatory bodies such as the Human Fertilisation and Embryology Authority (HFEA) and the American Society for Reproductive Medicine (ASRM) were included. Priority was given to studies reporting clinical outcomes, safety data, and ethical considerations. Both experimental and observational studies were reviewed, along with hypothesis and theory papers that introduced emerging technologies such as sperm epigenetic profiling and machine learning applications in ART. Data were synthesized qualitatively, focusing on clinical implications, strengths, innovations, ethical challenges, and future perspectives.

DISCUSSION

Evaluation of assisted reproductive technology (ART) outcomes indicates that overall live-birth rates remain moderate and are highly influenced by maternal age, embryo developmental stage, cycle type, and underlying infertility etiology [20]. The adoption of frozen-thawed embryo transfer (FET) and blastocyst-stage embryo transfer has increased markedly, with meta-analyses demonstrating comparable or superior live-birth outcomes relative to fresh transfers [21,22]. Consistent with national registry trends, our findings show that women under 35 years of age, those with high-quality embryos, and those with fewer infertility factors achieved the highest success rates. Furthermore, FET cycles were associated with improved neonatal outcomes—including reduced preterm delivery risk and higher average birth weight—when compared with fresh transfers [23,24]. These results underscore that ART success is determined by a constellation of interacting factors, reinforcing the need for precise patient selection and individualized treatment protocols. Clinically, the heterogeneity of ART outcomes highlights the necessity of personalized counselling: younger women and those with favorable prognostic factors derive the greatest benefit. The advantages observed with blastocyst transfer and FET likely reflect improved embryo selection accuracy and enhanced endometrial receptivity, as ovarian stimulation in fresh cycles may adversely impact the uterine environment [22,25]. Maternal age remains the most powerful predictor of ART success—implantation and live-birth rates decline sharply, while miscarriage and obstetric complications rise significantly in women aged ≥40 years [20,26]. Likewise, infertility diagnoses such as diminished ovarian reserve, endometriosis, and severe male-factor infertility are associated with lower oocyte yields and reduced embryo quality. The broader adoption of elective single-embryo transfer (SET) and blastocyst culture has reduced multiple gestations and associated complications [27,28]. Overall, our findings align with extensive registry data reporting live-birth per-transfer rates of approximately 20–25% globally, with superior outcomes in younger populations [20]. The increasing prevalence of FET and SET, and their correlation with improved perinatal and neonatal health, are well documented [21,23]. While some earlier studies favored fresh transfers, large contemporary meta-analyses now indicate that FET provides equivalent or even superior results in selected patients when endometrial conditions are optimized [25]. The cumulative evidence supports an evolving paradigm in ART—one that emphasizes individualized protocols, strategic use of FET and SET, precise consideration of embryo stage and infertility diagnosis, and comprehensive counselling on age-related prognostic factors. Recent technological innovations have profoundly transformed ART practice, enhancing both clinical success and patient experience. Artificial intelligence (AI) has emerged as a major driver of this transformation. AI-based systems such as ERICA and Life Whisperer use deep-learning algorithms to analyze embryo morphokinetics, achieving implantation prediction accuracies of 70–85% [2]. These tools enable more objective embryo selection, reduce inter-observer variability, and shorten time to pregnancy. In severe male-factor infertility, particularly non-obstructive azoospermia, the STAR system developed at Columbia University employs AI-guided imaging to identify rare motile sperm, potentially reducing the need for invasive surgical retrieval procedures like micro-TESE [3]. Machine-learning frameworks are also being integrated into predictive models that combine male and female reproductive parameters—including sperm epigenetics, paternal age, and lifestyle factors—with clinical markers such as anti-Müllerian hormone (AMH) and antral follicle count (AFC). These multi-dimensional models improve diagnostic accuracy and facilitate personalized ovarian stimulation regimens. AI-driven dosing algorithms have been linked to reduced ovarian hyperstimulation syndrome (OHSS) incidence and more cost-effective treatment cycles [2,9]. Another milestone in reproductive medicine is mitochondrial replacement therapy (MRT), developed to prevent transmission of pathogenic mitochondrial DNA (mtDNA) mutations. Techniques such as maternal spindle transfer (MST) and pronuclear transfer (PNT) replace defective mitochondria with donor mitochondria, enabling women with mtDNA disorders to have genetically related, disease-free children. The first clinical cohort in the United Kingdom reported eight healthy live births from 22 treatment cycles, with no evidence of mitochondrial disease to date [4]. MRT represents a paradigm shift from managing mitochondrial disorders to preventing them at the germline level, preserving the genetic connection between parent and child. Parallel progress in fertility preservation has produced promising experimental techniques such as artificial ovaries, constructed using bioengineered scaffolds seeded with primordial follicles, potentially offering cancer survivors fertility restoration without reintroducing malignant cells [6]. In vitro gametogenesis (IVG)—the derivation of gametes from somatic or induced pluripotent stem cells—could eventually provide reproductive options for individuals with premature ovarian insufficiency, prepubertal cancer survivors, and same-sex couples seeking genetically related offspring [1]. Collectively, these advances are steering ART toward a model of precision reproductive medicine, in which treatment is increasingly data-driven, personalized, and efficient. Nonetheless, rigorous long-term monitoring of children conceived through emerging technologies such as MRT and IVG remains critical to ensure safety and ethical integrity before widespread adoption [4]. The acceleration of ART innovation has intensified ethical and social debates surrounding its application. Germline modification through MRT, while therapeutic in intent, introduces heritable genetic changes, prompting concern over unforeseen transgenerational effects and potential progression toward non-medical germline editing or “designer genetics” [5]. Regulatory oversight, such as that provided by the UK’s Human Fertilisation and Embryology Authority (HFEA), has set a precedent by authorizing MRT only for serious mitochondrial diseases under strict conditions. However, global regulatory inconsistency fosters “reproductive tourism,” as patients seek treatment in countries with more permissive frameworks. Donor welfare and informed consent remain paramount ethical priorities. Oocyte donation entails physical and psychological risks—including OHSS and long-term uncertainty—necessitating comprehensive counselling, fair compensation, and transparent disclosure of risks [8]. Moreover, the psychosocial implications for MRT-conceived individuals, who inherit DNA from three genetic contributors, require long-term psychological follow-up and societal engagement. AI integration also presents unique ethical challenges. Many AI systems function as opaque “black boxes,” offering limited transparency into their decision-making processes, which may undermine patient autonomy and clinician accountability. Bias in algorithmic training datasets could also lead to inequitable outcomes across different populations. Furthermore, AI applications rely on extensive datasets containing sensitive reproductive and genetic information, raising serious concerns about data security and privacy [7]. Equity in ART access represents another pressing issue. Advanced interventions such as MRT and AI-based embryo selection remain cost-prohibitive and geographically restricted, widening disparities between affluent and resource-limited populations. As infertility is recognized as a disease by the World Health Organization, ethical frameworks must address distributive justice by promoting equitable access through subsidized or tiered pricing models. Finally, frontier technologies like IVG and artificial ovaries invite profound ethical reflection. IVG could enable genetic parenthood for same-sex couples, single individuals, and post-menopausal women, challenging conventional notions of family and reproduction. These possibilities necessitate proactive ethical deliberation, legislative clarity, and public dialogue to ensure socially responsible integration. The future of ART is moving toward precision, regeneration, and automation. In the near term, explainable AI (XAI) systems are expected to replace current opaque models, offering transparent decision-support for clinicians and improving patient trust. Large, multi-center validation studies are essential to ensure generalizability and minimize bias in these systems. MRT is anticipated to broaden its clinical indications as long-term safety data accumulate [5]. Refinements to minimize residual heteroplasmy and protocol standardization for MST and PNT are ongoing. Regulatory authorities must enforce lifelong follow-up of MRT offspring through centralized registries to detect potential delayed or transgenerational effects. Regenerative approaches such as artificial ovaries and IVG hold the potential to redefine fertility preservation. Artificial ovaries are expected to enter clinical trials for cancer survivors within the next decade, pending resolution of challenges like vascular integration and complete elimination of malignant contamination. IVG remains preclinical but could revolutionize reproductive options by generating functional gametes from somatic or stem cells—offering solutions for infertility, same-sex parenthood, and delayed reproduction. However, IVG must undergo rigorous genomic and epigenetic safety testing before clinical application. Non-invasive embryo assessment represents another exciting frontier. Metabolomic and genomic analysis of spent culture media could eventually replace invasive trophectoderm biopsies in preimplantation genetic testing (PGT-A), minimizing embryo manipulation and enabling longitudinal viability assessment. In addition, growing evidence implicates the endometrial and vaginal microbiome in implantation success, suggesting that future ART protocols may integrate microbiome modulation strategies—such as probiotics or targeted antibiotics—to improve outcomes. Ultimately, the establishment of global ART registries and collaborative research networks will be crucial to monitor long-term health outcomes of ART-conceived individuals, particularly those born through novel interventions like MRT and IVG. Harmonized international standards will promote ethical consistency, data transparency, and equitable access. Public education and policy frameworks must evolve in parallel, ensuring that ART innovation proceeds safely, ethically, and inclusively.

Key Innovations and Clinical Implications