FAQ
Frequently Asked Questions
What is OpinioAI?
OpinioAI is a research platform that uses AI language models to help researchers source relevant insights, data, opinions in a simple and scalable way.
Previously researchers would rely on survey, polls and interviews to get the data for their research. With modern AI, these legacy methods can be augmented, if not totally replaced.
How does OpinioAI work?
Researchers with Brigham Young University in their study "Out of One, Many: Using Language Models to Simulate Human Samples" discovered that GPT-3 has biases that are "fine-grained and demographically correlated, meaning that proper conditioning will cause it to accurately emulate response distributions from a wide variety of human subgroups."
In other words, you can simulate people using GPT-3 and they might respond with uncanny similarity to real people in real life. They called this approach "silicon sampling", and we're using a similar approach to provide insights and opinions at scale in 2 different ways.
First, the simpler one, where we apply the persona traits (personal and business) over a vanilla GPT-3 models to get relevant and helpful opinions and insights.
Secondly, we fine-tune the GPT-3 models with a relevant and arbitrary dataset to build a bespoke model that can be used for discovering insights and nuances on a much deeper and more precise level. This approach is how we get to the silicon samples.
What is Synthetic sampling?
Our definition of synthetic sampling is synonymous with silicon sampling. In essence it is a process of fine tuning and conditioning the GPT3 model so it "is able to produce outputs biased both toward and against specific groups and perspectives in ways that strongly correspond with human response patterns along fine-grained demographic axes. In other words, these language models do not contain just one bias, but many."
We replicate this process for any bespoke dataset and enable researchers to reach the same results as if they were polling or surveying real humans.