They're not claiming Google will hand out a user's biometrics to insurance companies or other parties. They're claiming Google will offer biometrically-based signal as an ad discrimination channel.
At that point, they don't have to hand out the data to allow insurance companies to discriminate; insurance companies could set up targeted ads for Cadillac insurance programs on users having XYZ biometrics and those ads only run for users that are showing they're low health risks. So it's soft market discrimination; only the extra-healthy get explicitly told about new insurance plan A, and the self-select group signs up, possibly making A cheaper for the insurer to implement even if they don't do any additional screening.
I think the theory has holes, but it's not nonsense. And it's unclear to me that it demands government regulation (but my bias is American-style law; I don't know how European legal eyes see the story I just laid out).