It's obvious (to me) that there are always humans behind a technology to verify and improve it. So voice recognition stuffs make no difference, yes, some people are hearing personal assistants records... but shhhh... Remember when Google was hiring pigeons to rank sites ? ... :)

Not sure if you read that article, but this is going to outside contractors.

Google has now made a statement on this over the data leak.

We just learned that one of these language reviewers has violated our data security policies by leaking confidential Dutch audio data. Our Security and Privacy Response teams have been activated on this issue, are investigating, and we will take action. We are conducting a full review of our safeguards in this space to prevent misconduct like this from happening again.

[blog.google...]

You would be surprised to know how much stuff are outsourced by big companies :)

"a small set of queries" - around 0.2% of all audio snippets

Small? What is small about 0.2%? That is 2 in 1000 queries. Roughly estimated from a Tech Crunch article (linked below) about 10M homes have a device in the US. Being conservative, let's use 1 query per day per device to see how small small really is.

Some math:
10M * 1 * 365 = 3,650,000,000 queries per year.
3.65B * 0.2% = 7,300,000 queries transcribed.
Yup that is a small number!

Now let us view this from a slightly different perspective. What is the probability that one of you searches was in the small set.
More math:
2 in 1000 = 1:500
500/365 = 1.37
1.37 * 12 month = 16.4 month
At one query per day one would expect that after 16.4 month at least one query would be included in that "small" set. If one uses the device more frequently one would reach that point of inclusion far sooner.

I've got nothing to hide... (head in sand)

Tech crunch article: [techcrunch.com...]

I see the picture. People are getting steamrolled. Not just by Google. All the big companies. It's like news story after news story. This is creating? Complacency. The biggest companies in the world are pushing the envelope and as it seems, the outrage or fight against data collection and data use is becoming background noise. Overrun. That's how I look at this. It still blows my mind people are stupid enough to use a box like this in their house, let alone pay for the device. Clearly people don't realize how valuable their data and implant is. The boxes are an implant yet people just don't realize it. They would give them away for free but then it would be obvious the desperation. Google has to beat out Amazon. Amazon has to beat out Google. The end game is connecting you by the vein to whatever revenue generation they can create out of being inside your house and life. How very stupid this is.