The dark side of ChatGPT 🌒🎭

Cliquez ici pour lire en français

ChatGPT, the successful conversational robot from OpenAI, which is causing such a stir on the web, does not only have its shiny sides. What you need to know, and what is currently causing a scandal, is that several Kenyan hands have worked quite hard to achieve this result.

The working conditions of the Kenyan teams 😱

So, you have to know that a huge work has been done by a team of Kenyans so that this artificial intelligence (AI) can provide us an appropriate content while respecting a certain ethics. The goal was to make sure that ChatGPT would not drown in the dark side of the web. It should be noted that its design is based on the contents present on the web until 2021.

It was a rather tedious job in which they had to deconstruct hundreds of descriptions selected by OpenAI. The most difficult part was that they contained details of violence, sexual abuse and hatred. So they had to be read and sorted out.

Being exposed to these different and quite violent contents on a daily basis, it was of course necessary to have recourse to « wellness » counselors who could help the employees to ensure their mental balance. After several investigations, it turned out that it was almost impossible for employees to have sessions with these counselors due to the high workload of the employees, but not only that, his specialties were unavailable most of the time.

Another aspect: the salary. For such a stressful job that can destabilize the mental balance of the employees, they were paid a very low salary in a difficult context (no benefits or vacations).

How was the connection made between OpenAI and its Kenyan employees ?  🔎

In order to sort through the content of ChatGPT, OpenAI hired Sama, an American company based in California, which hired several Kenyan employees for this purpose. It also describes itself as an « ethical » company, which is surprising considering the treatment of its Kenyan employees, who are paid two dollars an hour with no benefits.

In February 2022, a similar situation occurred with Facebook. Sama was accused of exploiting Kenyan moderators on behalf of Meta.

In the end, and in view of the problems with Meta, Sama and OpenAI had to break their contract some time before it expired. This, after discovering the violence of the content (texts and images) sent by OpenAI.

What do you think about this dark side of ChatGPT? Tell us all about it in comments.

Source : clubic.com