The reported 'AI blood test' in India: what it actually measures

Key takeaways

• The technology reported is photoplethysmography (PPG), which reads skin blood flow via light, not an actual blood draw or lab analysis.
• Calling it a 'blood test' overstates what camera-based vital-sign estimation does.
• Specific accuracy figures and the full list of measurable parameters were not independently verified for this article.
• PPG-based heart rate and oxygen estimation is an established technique used in fitness wearables; the claims beyond that need more scrutiny.
• This is medical and health technology reporting, which is off this site's core IT and cybersecurity training subjects.

What was reported?

Indian media reported a facial-scan health assessment tool, described as using a smartphone or tablet camera to estimate certain vital signs within about a minute, without a needle or blood draw. It was reportedly developed by a Hyderabad-based startup and piloted at a hospital.

Is it really a "blood test"?

No, not in the way that term is normally used. A blood test involves drawing blood and analysing it, directly or in a lab, for its actual chemical and cellular content. What is described here is photoplethysmography (PPG): a camera observes subtle changes in light reflected from skin, caused by blood flow with each heartbeat, and software estimates signals like heart rate and blood oxygen saturation from that pattern. This is a real and established technique, used for years in fitness trackers and smartwatches for heart rate and oxygen estimation. Calling the broader output a "blood test" overstates what the underlying measurement actually is.

What can PPG reasonably estimate?

Established, widely used PPG applications include:

• Heart rate.
• An estimate of blood oxygen saturation (SpO2), with known accuracy limitations compared to clinical pulse oximeters.
• Some estimate of heart rate variability.

What claims need more scrutiny?

Reports describing this specific tool mentioned a wider list, including blood pressure, HbA1c (used to estimate diabetes risk) and a "stress index," derived from a short facial scan. These are more ambitious claims than standard PPG use, and this article does not assert their accuracy, because independent, peer-reviewed validation was not available to verify at the time of writing. Readers and especially anyone making a health decision should ask for the actual validation data and regulatory status (such as any medical device clearance) before trusting such results over conventional testing.

Should this replace a real blood test?

No. Readers should not treat an estimate like this as a replacement for a laboratory blood test, especially for a diagnosis such as diabetes, which relies on validated blood chemistry testing. A screening tool, even an accurate one, is not the same as a diagnostic test, and the distinction matters for real health decisions.

Why is this article cautious rather than promotional?

Health claims deserve the same evidence standard as any other factual claim: a named source, published validation and, ideally, independent review. This article avoids repeating specific accuracy percentages or a full parameter list because they could not be independently verified, in line with this site's policy of not asserting figures it cannot check.

Does this fit an IT and cybersecurity training site?

Only loosely, through the AI angle. The core value for WebAsha's readers is understanding how AI and sensor-based estimation works, which connects to broader AI literacy, rather than the medical claims. The National Vulnerability Database and similar standards bodies are better references for the AI-security side of this topic than any health claim made here. An editor should decide whether to keep, redirect or retire this article, since it sits outside the site's core subjects of Linux, cybersecurity, cloud and DevOps training.

What should a careful reader do?

• Treat "AI blood test" framing in headlines with scepticism; check what is actually being measured.
• Ask for regulatory approval status and validation studies before trusting a new health device.
• Use established, validated devices for anything that affects a real medical decision.
• Consult a doctor for diagnosis, not a facial scan app.

Next steps

For a technical look at how AI is applied in diagnostics generally, see AI in healthcare diagnostics. If you want to build AI skills, see our AI Engineering course.

Related reading

• The Ascension Ransomware Attack | A Wake-Up Call for Healthcare Cybersecurity
• AI vs Hackers | How Indian Startups Are Using Artificial Intelligence to Beat Cybercrime in 2026