In Google and everyone else's case, I would argue it's a bit different. Google remunerates me with the best search engine in the world, excellent free email, etc. Other services improve the product, which I (generally) reap the benefits of.
On the other hand, your genetics are what they are -- I don't believe 23andMe is sending every participant a coupon for this new drug, although I might be wrong. Instead, all you get is a small sample of the 600k SNPs available, some shaky genetic inferences, and tabloid level ethnic data. You give up valuable data with huge upside for analysis of questionable integrity and limited actionability.
From the Privacy Highlight:
"We will not share your data with any public databases "
From the full policy:
"We may share some or all of your Personal Information with other companies under common ownership or control of 23andMe, which may include our subsidiaries, our corporate parent, or any other subsidiaries owned by our corporate parent in order to provide you better service and improve user experience"
I'm no lawyer, but it seems like the key word is public. Considering 23andMe reserves the right to share this data with its family of companies, and I'm unable to opt-out, it seems like all research is not totally avoidable. I won't comment any more about the validity of the science, it's not my field. However, I couldn't find any data about the error rate of the tests, which is surely non-zero. I'd love to know what SNPs they choose to evaluate, how they collect the phenotype data (the most valuable part of all!), etc.
In short, consumers give up huge troves of info, and pay for the privilege of doing so. What's more, the actual business model is turning this into drugs, of which they receive nothing in return. Most people expect a fun way to see they're 4% $ethnicity, not realizing the value in the data they freely give up to a private company.
I am no fan.