OpenAI is facing renewed criticism over claims that workers in Kenya were used to help train ChatGPT to detect and filter harmful content.
The criticism centers on the role of human workers in teaching artificial intelligence systems how to sort through data and identify unsafe material.
Technology companies have long used human input to train AI systems.
The provided text cites Google as an example, saying the company used CAPTCHA to help train Waymo systems to identify objects such as stop signs and traffic lights.
The text then claims OpenAI used a more harmful approach while training ChatGPT to filter dangerous content.
According to the claims, an OpenAI contractor hired workers in Kenya to review disturbing online material throughout the day.
The content reportedly included violence and sexual abuse, which workers reviewed to help train ChatGPT to recognize and filter unsafe material.
Time Magazine was cited as reporting that the workers were paid between $1.32 and $2 per hour, depending on “seniority and performance.”
The criticism argues that safe AI systems were built through the work of underpaid and traumatized human workers.
The text claims that thousands of workers were affected by this process.
It also claims that OpenAI’s president is Trump’s biggest donor and that the company is lobbying for legal immunity if ChatGPT causes deaths.
The text says there is a growing international movement to boycott ChatGPT.
It also directs people to follow the account QuitGPT to learn more about the campaign.
📢 For the latest Tech & Telecom news, videos and analysis join ProPakistani's WhatsApp Group now!
Follow ProPakistani on Google News & scroll through your favourite content faster!
Shares
