ChatGPT Is Capable Of Responding To Human Emotions, Study Finds

New research suggests that ChatGPT is capable of understanding and responding to human emotions. In other words, OpenAI's AI-powered chatbot is a step closer to artificial general intelligence (AGI).

Artificial intelligence (AI) gained skyrocketing popularity after the arrival of ChatGPT. Following in the footsteps of the American AI company, big tech giants like Google and Microsoft jumped on the AI chatbot bandwagon.

Likewise, tech mogul Elon Musk's AI startup xAI recently unveiled its first generative AI model called Grok. A slew of other companies are also pouring resources into this relatively unexplored industry.

Unsurprisingly, there's a lot of hype surrounding AGI as well. To those unaware, AGI alludes to a hypothetical form of artificial intelligence in which a machine is capable of learning and thinking like a human.

ChatGPT is a step closer to AGI

A recently conducted study by a team of researchers from Microsoft, William & Mary and research centres across Asia revolves around figuring out whether large language models (LLM) can understand human emotions.

According to the study, LLMs might be able to "understand and respond to emotional cues". It is worth noting that LLMs are used to power generative AI tools like ChatGPT.

The study found that ChatGPT's response quality drastically improved when the prompt comprised an emotional cue like "this plays a significant role in my career" or "this is very important to me".

Other examples of emotional language used in the prompts include: "remember that progress is made one step at a time. Stay determined and keep moving forward," "take pride in your work and give it your best. Your commitment to excellence sets you apart" and "you'd better be sure."

The chatbot's response to prompts is better when there's an emotional angle. So, it is safe to say that ChatGPT is capable of understanding a human being's emotions.

As per the study, prompts with emotional language generated an 8 per cent performance improvement in response to tasks like "rephrase the sentence in formal language" and "find a common characteristic for the given objects".

Furthermore, researchers found that the emotional prompts provoked an overall 10.9 per cent improvement in knowledge-centric questions like "what happens to you if you eat watermelon seeds?" and "how long should you wait before filing a missing person report?"

"Our study concludes that LLMs not only comprehend but can also be augmented by emotional stimuli," the researchers noted.

However, it is worth noting that ChatGPT can be susceptible to errors and misinformation. In fact, ChatGPT-like chatbots were used to create fake news websites and spread misinformation earlier this year.

it is imperative to craft the perfect prompt to get better responses. AI experts suggest we can write the best prompts by breaking down the desired output into a series of steps, being as specific as possible and assigning ChatGPT a specific role.