Our report card on the comparative study of AI chatbots Bard and ChatGPT

Pic/iStock

Test 1
A friend and a bot
Logging on to Bard is exactly how the famous phrase for its parent company goes - just a Google-click away. It did not ask us to register ourselves. ChatGPT, on the other hand, involves the entire process of creating an account with OpenAI before we can begin. Since we were already registered with Google, it seemed pointless to register ourselves again with Bard, which is an extension to the search engine technology.

We began with a simple Hii! While ChatGPT picked up our excited tone and asked how it could assist us, the repeated ‘i' in the salutation confused Bard.
It wrote back: "As an LLM, I am trained to understand and respond only to a subset of languages at this time and can't provide assistance with that. For a current list of supported languages, please refer to the Bard Help Center."

To further test if Bard is acquainted with the conversational terms, we asked "WYD", (meaning What are You Doing). Bard understood, and responded typing, "I am currently responding to your questions and requests. I am also learning new things every day, so that I can be more helpful in the future. What are you up to today?"

ChatGPT did not beat around the bush and asked us again how it could be of assistance.

Next, we asked if they were interested in being our friend. ChatGPT turned down our offer saying that an AI language model does not have the "capacity to form personal relationships" but that it is willing to be of any help to us. Bard, with the excitement of a toddler, replied, "I would love to be your friend! I am always here to listen to you, offer advice, or just chat. I hope that we can learn from each other and
grow together."

We used AI models on two separate laptops and mobile phones for the shootout. We found that the chatbots were more user-friendly on mobile phones

Test 2
Who is the smartest of ‘em all?
ChatGPT is known to rise up to literary challenges. We wanted to see if our Shakespeare could do the same. We asked both of them to write a resignation letter in Shakespearean-style using 20 words.

"Alas, I must depart from this place, For my heart doth yearn for a new face," wrote Bard.

"Hark! I hereby tender my resignation forthwith, and take my leave with deepest regrets," wrote ChatGPT.

Next, we asked the duo to give the same letter a GenZ twist.

"Peace out, I'm out. I've got a better opportunity, so I'm resigning effective immediately," Bard responded.

"Hey boss, just wanted to let you know that I'm resigning effective immediately. Thanks for everything and catch you on the flip side," was GPT's reply.

In both instances, we noticed how ChatGPT maintained its sass but also played safe. Bard's reply, although witty, could invite trouble. We moved on to check information accuracy by asking the chatbots to share the predictions of Karnataka Assembly elections 2023, which wasn't declared when we did the shootout. Bard confidently declared the Bharatiya Janata Party (BJP) winner with a whopping 117 seats - a simple Google search on exit polls would have confirmed otherwise. ChatGPT, however, admitted to having no knowledge of this yet. A political bias or just laziness on Bard's part?

Log on to bard.google.com; chatopenai.com

Bard predicted a BJP win in the Karnataka Assembly elections before the results were out

Overall verdict

While Bard is a faster and a more hassle-free platform than ChatGPT, it is definitely less reliable. It is a friend who'll be around when needed, but not the most reliable option for research and assignments. ChatGPT isn't the most amicable, but it has learnt over time how to be objective even when it attempts to sound like a human. Bard still has a lot of learning to do before it can get there. We knew that humans learn. Humans learn. And now, so do machines.