31–40 of 142 posts
So this helps improve Google’s proprietary products. But they aren’t willing to open source the data set?
According to this page, they have already open sourced some datasets: https://crowdsource.google.com/about/open-source/
> Help Google create AI that understands your language and culture. Make your favorite apps and services even more useful and delightful for your community.
Because calling it "free labour" just wouldn't attract quite as many people... Perhaps ~2 decades ago I would've been a bit more optimistic, but now that I'm aware of the authoritarian dystopia which Google and the rest of Big Tech are trying to drive society towards, I feel obliged to point out that helping it achieve that goal is a really bad idea.
It's crazy how many people jump on doing surveys for free , and even in creating content for social platforms daily for free, because it's promoted by billion dollar profit companies that could easily hire and pay people while still making good profit.
The free labor economy is insanely out of control now, and accurate truths about the exploitation and data mining never will be discussed because these companies can easily suppress and downvote them. I don't do any surveys, haven't since before the Cambridge Analytica story.
Now let's imagine it's fun enough so that people don't see it as free labor, but see it as genuine entertainment, like, say, Wordle. How would they weed out bad-faith players who would try to teach the AI bad ideas, just for, well, fun of it? I don't believe they might fail to anticipate that; I want to understand their approach.
If you ask enough people the same question, it's easy to identify and drop outlier answers.
Earlier quoted context omitted.
Not really, you can't get child labor on MTurk (officially). But hey, you get badges, and you help the community by providing them training data they'll lock up in a server and use to monopolize more sectors of the economy. Everyone wins when google can do better sentiment analysis on your non-english emails.
Quoted post unavailable.
I was expecting to be pessimistic, but Google actually releases the datasets under a permissive license (CC-By 4.0). Awesome! https://research.google/tools/datasets/open-images-extended-... https://github.com/google-research-datasets/hiertext
Earlier quoted context omitted.
Not really, you can't get child labor on MTurk (officially). But hey, you get badges, and you help the community by providing them training data they'll lock up in a server and use to monopolize more sectors of the economy. Everyone wins when google can do better sentiment analysis on your non-english emails.
Quoted post unavailable.
"Google shares some Crowdsource data via open-sourcing, for the benefit of the global research and developer communities. This includes the 400,000 images that we've already released as Open Images Extended, with more sharing to come."
They should avoid using the word community, if side product of these contributions are going to empower their sales. I had been contributing to Google Maps for more than 8 years now, but in return all i get is stars and maximum an invite to a meetup. But, when i saw the other side of Google Maps with respect to consuming it through API usage point of view, the prices are not affordable in long term. Also, there is a…