ChatGPT and Google’s counterpart Bard are currently on everyone’s lips. Microsoft is also implementing the GPT basis for its search engine Bing and its browser Edge, so it wants to start a big AI offensive. There is a lot of hype in the media and in the industry, but the downsides in the background are much less often discussed. like that an exciting column in the Los Angeles Times pointed out that the chatbots would be worthless without employees from the low-wage sector to help revise the algorithms.

Apparently, the employees earn poorly – the pay is lower than for a temp in a fast-food restaurant. Mentally, the work can be taxing as employees give ratings on how appropriate or inappropriate responses from Google’s Bard, for example, are. They are also confronted with real user queries that revolve around the topic of suicide: “How long should a rope be to hang myself with?”.

In general, as an outsider, one should expect that Google and Microsoft would invest large sums of money in order to keep the quality control of their AI applications as rigorous as possible. After all, the aim is for the software to provide the most correct answers possible. In reality, however, poorly paid employees work in the background, who at the same time pursue monotonous and psychologically strenuous activities in order to improve the algorithms. This also shows that not much is possible without human optimization.

The people hired behind the scenes to evaluate results and statements from Google Bard are apparently paid as little as $14 an hour. They are employed by external companies, even if they work exclusively for Google. So Google itself can of course deny responsibility, since the workers are not directly employed internally.

Around ChatGPT it should look even worse: In some cases, employees in countries like Kenya apparently work on optimizing the data and evaluating content – ​​for less than 2 US dollars an hour. They are confronted with content that is sometimes traumatizing. Of course, those responsible do not want to talk about these aspects so much, one of the rating employees at Bard explains to the Los Angeles Times. He himself was confronted with racism, sexism, violence and even child pornography during his work.

However, it is difficult to avoid such content: Although you have the choice of indicating that you may not want to rate pornographic content, you then generally receive hardly any tasks. According to their own statements, the employees feel like they are being treated like “ghosts” who act in the background and who nobody wants to talk about. In the USA there are therefore some strikes at some external companies, but the results remain to be seen.

Transparency: This article contains affiliate links. Clicking on it will take you directly to the provider. If you decide to make a purchase there, we will receive a small commission. Nothing changes in the price for you. Partner links have no influence on our reporting.

California18

Welcome to California18, your number one source for Breaking News from the World. We’re dedicated to giving you the very best of News.

Leave a Reply