Ever wonder what makes one AI chatbot seem almost like it’s a person speaking/responding to your specific questions and another one seems like nothing more than a simple chatbot with pre-programmed responses to set questions?
In most cases, the answer is AI training.
More specifically, it’s what’s known in the AI industry as data annotation that gives certain AI systems what feels like a spark of humanity.
It involves labeling or tagging raw data with relevant metadata to “teach” AI and machine learning models how to understand and interpret the data. It helps the AI recognize patterns, objects, sentiments, and much more in ways that humans do without realizing they’re doing it.
Before a generative AI or a chatbot is released into the wild, behind the scenes, people have trained the AI by going through a series of generated responses to prompts and grading and scoring them, giving hints and tips that the AI then can factor into future generative responses to prompts.
For the people, it’s a chance to make some money by doing short tasks on their schedule. It can be gig economy work, but it also has opportunities for scaling up even as much as replacing full-time work.
January’s React Denver talk will be on AI training. Jenni Grubbs has worked as a tasker and reviewer for multiple data annotation platforms over the last three years. She has helped train various AI systems on a wide variety of projects and will share insights into how it works, how to get started, what can be gained from starting a side hustle in data annotation, the downsides, and more.
About Jenni Grubbs
Jenni Grubbs has worked as a tasker and reviewer for multiple data annotation platforms over the last three years. She has helped train various AI systems on a wide variety of projects and will share insights into how it works, how to get started, what can be gained from starting a side hustle in data annotation, the downsides, and more.
Hosted by Netlify