TechCrunch reports Mantis Biotech is building human digital twins

In short: TechCrunch reports that Mantis Biotech is building “digital twins” of humans using synthetic data to help address a lack of medical data.

What happened

TechCrunch says Mantis Biotech takes data from different sources and turns it into synthetic datasets. Synthetic data is computer-made data that is meant to look like real data, like a flight simulator that imitates real flying.

The goal, according to the report, is to help create “digital twins” of the human body. A digital twin is a digital copy used for testing ideas, like using a virtual crash-test dummy instead of a real person. TechCrunch describes these digital twins as representing anatomy, physiology, and behavior.

At the same time, the name “Mantis” also appears in AI research with a very different meaning. In published research, Mantis is described as a “foundation model” for time series classification. That means it is a general-purpose AI model trained to label patterns in data that changes over time, like heart rate, sleep signals, or motion sensor readings. That research says it uses synthetic time series for training to improve “zero-shot” performance, which is when a model tries a new task without special training for it.

Why it matters

Medical AI often runs into a basic problem: there is not enough usable real-world data, and patient privacy is a serious concern. If synthetic datasets are accurate enough, they could let researchers test ideas more safely and cheaply. But people should also watch for clear explanations of what data is used, how closely it matches reality, and whether a company is referring to a research model name or a separate product.

Source: TechCrunch AI