At this point, any request for information could potentially be used as training data. That includes things like captchas.
I recommend everyone have an extremely literal interpretation of “labor”. Unless you have tremendous insight into where your data is going and how it is being used (and perhaps even then), then assume any ask is ultimately an ask for unpaid labor.
Obviously you can’t avoid things like captchas, but you can avoid things like this.
Edit: and it should go without saying, but anything you upload to socials is probably automatic training data at this point. The best approach is simply not to engage with corporate social networks.
Though Lemmy is not corporately controlled, the information is publically accessible, so even this post is potential training data to be scraped. That is harder to avoid, lest we stop using the internet altogether, but at least avoiding the corpo routes is a good start.
I use an app called Be My Eyes to help the visually impaired.
You’ll get a random notification that a person needs your help. If you’re the first to respond, you’ll be paired up. Their phone camera is displayed on your screen, and you can talk to each other.
I always have a great experience when I use it.
They’re not winning over the blind again after they limited access to screen-reader-accessible apps.
I promise that I’m leaving this here in good humor, although it is also true.
OP did actually use it correctly tbh, a rare case
Context, please?
AI steali…I mean harvesting, pardon learning
I’m sure blind people are happy to have the models that are built with this data, and since both the image and the description are public facing, anyone can use them including open source.
Helping self-driving cars and drones.
