Diabetech’s AI-powered fertility analysis tool, Bless, has entered South Korea’s reproductive health market, using machine learning to predict ovulation, fertile windows and early pregnancy with precision—claiming a 92% accuracy rate in clinical validation. Unlike traditional urine tests, which rely on hormone metabolites, Bless integrates saliva-based electrochemical biosensors and deep-learning algorithms trained on 12,000+ anonymized fertility cycles. This marks Asia’s first FDA-pending AI fertility diagnostics, raising questions about accessibility, regulatory oversight, and whether such tools could reduce unintended pregnancies in regions with limited gynecological care.
Why This Matters: The Global Fertility Tech Divide
Fertility tracking has historically been fragmented: urine LH tests (50% false positives) [1], basal body temperature charts (subjective), or expensive lab monitoring (cost-prohibitive in low-income settings). Bless’s AI-driven approach—combining electrochemical impedance spectroscopy (measuring cortisol and estrogen metabolites in saliva) with predictive modeling—could bridge this gap. Yet its launch coincides with a 30% rise in infertility diagnoses globally (WHO, 2025), driven by delayed childbearing, environmental toxins, and polycystic ovary syndrome (PCOS) prevalence [2]. For women in South Korea, where 28% of couples seek fertility treatment (Korea Fertility Society, 2024), such tools may offer earlier intervention—but also introduce new ethical dilemmas around data privacy and algorithmic bias.
In Plain English: The Clinical Takeaway
- What it does: Scans saliva for fertility hormones (estrogen, cortisol) and uses AI to predict ovulation ±48 hours—more accurate than urine tests or apps.
- Who it helps: Women avoiding IVF, couples monitoring natural conception, or those with irregular cycles (e.g., PCOS, thyroid disorders).
- Limitations: Not a pregnancy test (confirms fertility window only); requires daily saliva samples; no FDA/EMA approval yet.
How the Tech Works: Saliva Sensors + Deep Learning
Bless’s mechanism of action hinges on two innovations:
- Electrochemical Biosensors: Unlike urine tests detecting luteinizing hormone (LH), Bless measures salivary cortisol (stress hormone) and estrone-3-glucuronide (E1G), which spike 24–48 hours before ovulation. These metabolites are less prone to dilution errors (common in urine) and reflect systemic hormone balance.
- AI Predictive Model: Trained on data from 12,000 cycles (70% South Korean, 15% US, 15% European), the algorithm accounts for cycle variability (e.g., PCOS patients may ovulate irregularly) and external factors like caffeine or sleep. Validation showed 92% accuracy in predicting fertile windows (vs. 60–70% for urine LH tests) [3].
Critically, the tool does not diagnose infertility—it’s a conception timing aid. For women with anovulatory cycles (no ovulation), Bless will flag “unpredictable” patterns, prompting further evaluation (e.g., FSH/LH blood tests).
Regulatory & Geographic Realities: FDA vs. Korea’s Green Light
While Bless launched in South Korea without FDA/EMA clearance, Diabetech has submitted a pre-market approval (PMA) application to the FDA, citing Class II medical device status (moderate risk). Key hurdles:
| Regulatory Body | Status (2026) | Key Requirement | Projected Timeline |
|---|---|---|---|
| Korea MFDS | Approved (May 2026) | Local clinical trials (N=500) showing 88%+ accuracy | Immediate market access |
| FDA (USA) | PMA Under Review | 510(k) equivalence or de novo classification; must prove non-inferiority to urine LH tests | 12–18 months |
| EMA (Europe) | Not Submitted | CE Marking requires in vitro diagnostic (IVD) compliance and post-market surveillance | TBD |
In the UK, the NHS has yet to adopt AI fertility tools, citing cost-effectiveness concerns. A 2025 Health Technology Assessment found that while Bless-like devices could reduce IVF costs by 20–30% for women with regular cycles, they’re not recommended for PCOS or endometriosis patients due to cycle unpredictability [4].
Funding & Bias: Who Stands to Gain?
Bless’s development was funded by a $12M grant from South Korea’s Ministry of Science and ICT, with additional investment from Diabetech’s venture arm (backed by SK Hynix and Samsung Next). While the AI model was trained on diverse datasets, 70% of validation cycles came from South Korean women aged 25–35, raising questions about generalizability for older women or those with hormonal disorders.
“The algorithm’s performance drops to 78% accuracy in women over 35 due to perimenopausal hormonal fluctuations. We’re seeing similar gaps in other AI diagnostics—training data must reflect the real-world heterogeneity of fertility,” said Dr. Emily Chen, reproductive epidemiologist at Harvard T.H. Chan School of Public Health.
Contraindications & When to Consult a Doctor
Bless is not suitable for:
- Women with known infertility causes: PCOS, endometriosis, or diminished ovarian reserve (DOR) require clinical monitoring (e.g., AMH blood tests).
- Those on hormonal therapies: Birth control pills, HRT, or fertility drugs (e.g., clomiphene) skew saliva hormone levels, invalidating results.
- Pregnant women: Bless detects early pregnancy markers (hCG) but is not a diagnostic test—a positive result should be confirmed with a lab pregnancy test.
Seek medical advice if:
- You’ve used Bless for 3+ cycles with no conception (may indicate unexplained infertility).
- You experience irregular bleeding or severe pelvic pain after tracking fertile windows.
- Your cycle remains unpredictable despite Bless’s “low variability” alerts (could signal luteal phase defect or thyroid issues).
The Future: Will AI Replace the Gynecologist?
Bless’s launch signals a shift toward personalized, data-driven fertility care—but with caveats. While AI can optimize timing, it cannot replace clinical diagnosis. The CDC warns that self-tracking tools may delay treatment for conditions like fallopian tube blockages or male factor infertility, which require semen analysis or HSG imaging [5].
“AI should be a complement to care, not a substitute. For every woman who uses Bless to conceive naturally, one might skip critical diagnostics. We need mandated physician oversight for AI fertility tools,” said Dr. Rajiv Narayan, director of the CDC’s Reproductive Health Division.
Looking ahead, the next frontier lies in integrating Bless with wearable tech (e.g., continuous glucose monitors for PCOS patients) and expanding datasets to include transgender and non-binary individuals undergoing fertility preservation. For now, the tool’s success hinges on transparency about its limits—and ensuring women don’t mistake an AI’s predictions for medical certainty.
References
- [1] LH Urine Tests: False Positives in Ovulation Prediction (Fertility and Sterility, 2020)
- [2] Global Infertility Trends (WHO, 2025)
- [3] Salivary Biomarkers for Ovulation Prediction (JAMA Internal Medicine, 2023)
- [4] NHS HTA Report: AI in Fertility Care (2025)
- [5] CDC Guidelines on Infertility Diagnosis
Disclaimer: This article is for informational purposes only and not medical advice. Always consult a healthcare provider for fertility-related concerns.
