Monday, June 12, 2023

Large language models need thousands of hours from human data annotators in every stage of AI development

So much for artificial intelligence (AI)! A look behind the hype!

"... the human labor behind fancy AI models
The secret to making AI chatbots sound smart and spew less toxic nonsense is to use a technique called reinforcement learning from human feedback [RLHF], which uses input from people to improve the model’s answers. ...
It relies on a small army of human data annotators who evaluate whether a string of text makes sense and sounds fluent and natural. ...
Even the most impressive AI chatbots require thousands of human work hours to behave in a way their creators want them to, and even then they do it unreliably. ...
Data annotators are involved in every stage of AI development, from training models to verifying their outputs to offering feedback that makes it possible to fine-tune a model after it has been launched. They are often forced to work at an incredibly rapid pace to meet high targets and tight deadlines ..."

We are all AI’s free data workers

No comments: