International
AIs are no match for human creativity, with people outperforming generative AIs such as ChatGPT and Claude on a creative word task, according to international research, which found the AIs did even worse when they were prompted to think like Shakespeare, DaVinci and even Oprah Winfrey. The researchers tested the creativity of people and AIs using a task that involved coming up with ten single-word nouns that were as different from each other as possible. The researchers found that people used a wider variety of words and had a broader vocab than the AIs. When the researchers tried to increase the creativity of AIs by prompting them to take on historically renowned creative figures, they found that their creativity scores generally got worse. The authors say this highlights the potential risks associated with over-reliance on these AIs for creative tasks, as it could inhibit innovative breakthroughs in society.
Note: Not all attachments are visible to the general public. Research URLs will go live after the embargo ends.
Sign In
