Live data from Hacker News

Kenyan moderators decry toll of training of AI

theguardian.com

21–30 of 91 posts

Re: Kenyan moderators decry toll of training of AI

#21
post #9

> Workers were paid between $1.46 and $3.74 an hour, according to a Sama spokesperson. I don't like how journalists never adjust the salary when doing those conversions. Yes $1.46 in the US sounds like the most abdject poverty, but Kenya is not the US. The minimum salary in Kenya is $1.05 so even their lower bound is 39% higher than the minimum salary with their upper bound being 358% higher. For job that while psych…

If you worked for Sama and was sent to work in Kenya, would you take a salary cut just because your cost of living has reduced? Pay people what they are worth, not based on where they choose to live.

Re: Kenyan moderators decry toll of training of AI

#22
Any form of repeated exposure to content or propaganda is bound to have a lasting impact on both our cognition as well as memory. This is how we are wired. The key is to ensure we are able to draw boundaries - a key reason why we are all talk about social feed / content moderation.

The challenge is when you can afford to pay and get someone else to do it - is it your problem and do you care if it affects someone else.

Do corporations bear responsibility for affecting people who they employ and put into severe conditions which they are aware have long lasting impact?

Re: Kenyan moderators decry toll of training of AI

#23

Based on the information in the article these complaints were raised after the contract was ended. Why not refuse the work or raise the complaint at the time? The article also mentions the work was worse than they thought, but indicates they were told what the work entailed before starting it. Also the primary person being quoted lost his wife and family, and blames this work on why he lost his wife. Anyone who has b…

It's activism, not journalism. The purpose is to persuade, not inform.

Re: Kenyan moderators decry toll of training of AI

#24
post #2

I wonder if executives of companies using mturk-like humans to moderate/classify vile content should be compelled to do the same moderation/classification work themselves for an hour a month. Of course it’s highly unlikely to actually compel them to do this through legislation, but it’s an interesting thought.

What would this achieve?

Either an application of empathy to the business, or a filter to allow only completely unempathetic people into leadership.

Re: Kenyan moderators decry toll of training of AI

#27
post #9

> Workers were paid between $1.46 and $3.74 an hour, according to a Sama spokesperson. I don't like how journalists never adjust the salary when doing those conversions. Yes $1.46 in the US sounds like the most abdject poverty, but Kenya is not the US. The minimum salary in Kenya is $1.05 so even their lower bound is 39% higher than the minimum salary with their upper bound being 358% higher. For job that while psych…

If you worked for Sama and was sent to work in Kenya, would you take a salary cut just because your cost of living has reduced? Pay people what they are worth, not based on where they choose to live.

> If you worked for Sama and was sent to work in Kenya, would you take a salary cut just because your cost of living has reduced?

These people weren't "sent to work in Kenya." They already live in Kenya, and are getting wages that are typical of jobs in Kenya.

Re: Kenyan moderators decry toll of training of AI

#28
post #2

I wonder if executives of companies using mturk-like humans to moderate/classify vile content should be compelled to do the same moderation/classification work themselves for an hour a month. Of course it’s highly unlikely to actually compel them to do this through legislation, but it’s an interesting thought.

Real imagery is one thing but it's hard for me to imagine being traumatized by classifying fictional content, especially text. Getting paid only a few dollars an hour might be painful, though.

Who knows what kind of texts people share though. Granted, you should not be a content moderator if you don't hold the same view as you expressed above, but I have a feeling a lot of these people weren't prepared at all for what people shared with OpenAI, and just got the job for the (relatively) high pay.

But humans can be fucking terrible beings. I could imagine they shared all types of content they had no idea would be read by another human, like confessions of past events and such.

Re: Kenyan moderators decry toll of training of AI

#29
post #3

Why were human data labelers involved in the creation of ChatGPT? I thought the approach was to let the neural network predict the next word in a large corpus of text?

I don't know the scope of OpenAIs data cleaning or RLHF efforts. But I'd guess this is a big part of what's made them the leader. It's one thing to train an LLM on a big text corpus, as researchers do. It's another to actually wade into it and do the curation to turn it into a real product. I know there are public instruction datasets out there, but my guess is OpenAI has put a lot of money into getting cleaner data and more supervised / feedback data. Doing that extra work is the kind if thing that wins when building a product.

Re: Kenyan moderators decry toll of training of AI

#30
post #9

> Workers were paid between $1.46 and $3.74 an hour, according to a Sama spokesperson. I don't like how journalists never adjust the salary when doing those conversions. Yes $1.46 in the US sounds like the most abdject poverty, but Kenya is not the US. The minimum salary in Kenya is $1.05 so even their lower bound is 39% higher than the minimum salary with their upper bound being 358% higher. For job that while psych…

I don't think anyone would dispute that these are not decent wages in Kenya. What is embarrassing is that the corporation seeks places where wages (relative to western wages) are this low (for what also amounts to the bullshit work).

Yes, it’s terrible that people in low income countries have jobs.
Post reply on HN