Has ChatGPT Superseded Human Intelligence? This New Study Provides A Unique Perspective
A leading number of high-profile research has gone on to evaluate AI and how it has managed to surpass the world of human creativity.
AI assistants like ChatGPT are being looked at more in detail now than ever, growing in their popularity. However, a new study is providing a new and unique perspective in terms of how the views are getting more nuanced as time goes by.
Researchers found that while AI does tend to excel in terms of deriving great interpretations linked to ambiguous images, humans are still looking at these responses as something that’s more creative.
Such results point to how human creativity continues to have that leading edge over AI in specific domains.
The fast developments and adoptions of such systems including ChatGPT, Llama, and even Bard continue to be the subject of great debate in terms of what impact they have on society and what their concepts linked to creativity are.
AI advancements and their ability to carry out so many tasks today are solely restricted to humans and there is so much concern about how humans could soon be displaced, how it could impact the education sector, and what ethical dimensions AI content brings forward.
It’s great to see how things were getting explored in this domain and how the advancements made in AI models could really produce outputs that made it seem so real and life-like.
The current understanding on this front has to do with how models fare well against the performance of humans. To be more specific, we’re interested in grabbing an understanding of how such changes would impact AI in a manner through which humans perceive creativity.
As a general rule of thumb, we’re intrigued by comprehending how AI could alter how humans take on creativity and the kind of impact such AI models would have socially.
Today, there’s a lot of talk on this front and how the excitement linked to concerns about AI surpassing human creativity is a trending sector. This is where we aim to investigate the current state of AI and how it fairs against the human mind.
To help gauge a better understanding on this front, it was interesting to see how the researchers explored various creativity capabilities linked to AI and humans through the use of Figural Interpretation Quest.
This involved interpreting some abstract figures. The test was rolled out to better comprehend divergent thinking entailing several different solutions concerning open-ended issues.
It was different with a convergent style of thinking that is more designed to focus on the likes of single and accurate responses.
The study featured 279 people who spoke English as their native language. On the other hand, we had two features examined, how creative it was and how flexible it was.
Coming to the results, the study’s authors found that ChatGPT-4 portrayed bigger than usual flexibility in producing interpretations when you compare that to humans. This proved how AI could give rise to more diverse ideas and display the chance to think divergently, no matter what context was involved.
Flexibility scores for AI models were greater than humans and that just goes to show how remarkable the language capabilities of LLMs are in today’s day and age.
Moving on to creativity, AI certainly fell short of its human counterparts. Humans were not only more creative, but their rich experiences proved that they could better interpret stimuli.
Human replies outperformed AI in terms of creativity and flexibility. AI could give rise to replies that were more narrow while humans gave rise to better creativity scores in general as they had a broader array of ideas.
Image: DIW-Aigen
Read next: From ‘Don’t Be Evil’ to Legal Battles: The Erosion of Google’s User-Centric Philosophy
