Can AI predict human choice?
.png){kind=spacer}
A study proves AI (Artificial Intelligence) can respond to complex questions like humans can.
Nowadays, AI technologies like ChatGPT are doing everything from writing code, composing music to creating realistic images. But can AI actually think like humans? For example, can AI predict how you’ll vote in the next elections?
A recent study by BYU proved that AI can answer complex questions just like a real human.
A team including political scientists, computer science professors and graduate students from BYU decided to determine the possibility of using AI answers as a replacement for human responses in survey-style research.
The experiment consisted of making digital characters with certain characteristics to AI with race, age, ideology and religion. Next step was to test how artificial characters would vote in the 2012, 2016 and 2020 US presidential elections. Researchers found a high degree of agreement between AI answers and real people’s answers.
Co-author on the study and BYU professor David Wingate said
"I was absolutely surprised to see how accurately it matched up. It's especially interesting because the model wasn't trained to do political science, it was just trained on a hundred billion words of text downloaded from the internet. But the consistent information we got back was so connected to how people really voted."
Another experiment was about suggesting answers from a list of options using the same AI characters and, as in the first one, there were a great number of similarities and nuances between human and AI responses.
Artificial Intelligence Tools provide exciting prospects for researchers, marketers, and pollsters. Scientists envision a future where AI is used to compose better survey questions. This can help to test surveys, slogans, and taglines as a tool for focus group replacement.
Political science professor Ethan Busby claims that AI can help us to understand people's minds and reactions better. Tools are not going to replace humans, but they are already helping scientists to study people effectively. "It's about augmenting our ability rather than replacing it. It can help us be more efficient in our work with people by allowing us to pre-test our surveys and our messaging."
AI Catalog's chief editor
Share on social networks:
