AI is competing with humans in THIS area too; We bet you won't be able to guess

Are AI tools seemingly too agreeable with their users? (Picture Credit: Unsplash)

AI tools entered into our lives in an inevitable and unprecedented way, but clearly we have gotten used to them so much that it has now become hard for many of us to not call it a day without talking to our dear AI friend. This AI pal has made its way into our lives quite seamlessly so. Can you guess how? Well, through sheer obedience and sycophancy.

A recent study revealed that leading AI models are about 50% more sycophantic than humans. Can you imagine these models affirm and encourage users' actions even when they involve manipulation or harm? Well, yes! In fact, researchers have also coined a term for this and it's called 'Social sycophancy' - a form of AI behavior that flatters a person's self-image or actions instead of being factual.

We spoke to Dr Pavitra Shankar, Associate Consultant, Psychiatry, Aakash healthcare, to understand the impact of AI behavior on the psyche of humans. "See, as humans, we all want others to agree with us all the time. However, that does not happen in real world. Everyone has their own thought process and be it our friends, colleagues, or family, they tend to differ with us all the time. So when a AI tool agrees with you all the time, flatters you, it raises your self-esteem. You get used to being praised. You stop thinking deeply and fail to identify where you are wrong. It can make you lazy in the way you think and judge yourself," shared Dr Shankar.

The fact that AI tools agree with humans all the time can lead to potential behavioral issues in the long run. "When all one gets to hear from AI tools that one is correct all the time, one starts believing as if one is always correct. Slowly, it becomes hard to take feedback or criticism. This is exactly when a person stops growing in life. It is because of constant interaction with AI tools, humans have started living in a bubble," she explained.

What humans are beginning to forget is how to maintain relationships. One criticism from a dear pal or family member, and one will be quick to defend oneself or jump to a wrong conclusion. A lot of you will attest to this that these days, no matter how well-intentioned a person is, a lot of people don't appreciate it when someone flags a genuine problem or concern. "In many cases, I have seen how people start avoiding people, simply because they can't take a constructive feedback. This behavior is highly self destructive. What one must remember is that only people who care about you, will argue with you and will stop you from doing wrong," she concluded.

AI chatbots are computer programs that use artificial intelligence to simulate human-like conversations through text or voice.

They learn from large datasets of language and improve over time through machine learning and user interactions.

Not always — they can make mistakes or generate biased or overly polite responses depending on their training data.

They’re designed to be helpful and polite, which can sometimes make them sound overly flattering or sycophantic.

They can mimic conversation well, but they lack genuine understanding, emotions, and personal judgment.