Google Bard vs. ChatGPT: Compared over 9 Prompts

Head to head, Google Bard vs. ChatGPT.

Generative AI has long come out of the buzzword stage.

What started with an OpenAI experiment (ChatGPT) quickly became the soon-to-be adopted industry standard, with Google and Microsoft following suit.

For those new, Microsoft has a huge stake in OpenAI, and it might have taken some pages from ChatGPT. So personally, ChatGPT doesn’t feel like a direct competitor of BingAI.

But what presents a true rival is Google Bard, which is still in beta as of this writing.

I’ll end this preface right away and will start giving a few prompts to Google Bard and ChatGPT. Ultimately, you’ll see who is better for your use case.

Let’s begin already!

Google Bard vs. ChatGPT 4

Via these prompts, I’ll try to illustrate the various skills of these AI engines. It’ll have something related to programming, tourism, drafting emails, storytelling, poetry, and a few everyday queries that we normally end up googling.

Fictional Story

Write a 50-word story about a gardener who secretly sells mangoes to Aliens.

Google’s Bard failed at keeping the word count, as its version was 175 words when the prompt clearly asked for a 50-word story. In addition, it didn’t name the gardener, and its story felt like a stolen abstract from somewhere around the internet.

Besides, it was full of cliche sentences without much creativity. Dull, boring, and unappealing.

ChatGPT, on the other hand, limited it to 47 words 👍 (under 50 words). Moreover, its version felt like you could attach it to your fictional novel, in contrast to Google Bard’s uninspiring Kindergarten attempt.

Finally, it indicated the trade details, like the gardener got rare minerals in exchange for mangoes. So this round clearly goes to ChatGPT with a mammoth margin. I mean, Bard was not even close.

ChatGPT: 👍 | Bard: 👎

Coding

Coding is another aspect where these generative AI tools are expected to shine. Maybe they can’t replace a skilled coder yet. Still, they can be worthy assistants capable of writing beginner code or preparing a shell for their human counterparts to build upon.

Please program a simple calculator in Python.

Next, I run this code in a Python compiler.

Bard’s version gave this:

And ChatGPT code wasn’t lacking either:

While we can adjust the calculators with subsequent prompts, Bard did better on this first attempt.

All things aside, Bard’s calculator prompts you to continue at the end of an operation, while for ChatGPT, you would’ve to run it again.

So, this round goes to the new entrant.

ChatGPT: 👎 | Bard: 👍

Drafting Email

Write a strongly worded first reminder email under 100 words to one of my clients about the non-payment of a $2500 invoice due last month.

I guess Bard has some kind of amateur functionality set in that we cannot alter.

It was a simple email, and I did mention it was a first reminder. Yet, it directly came to suspending services and threatened legal action, which was absolutely unnecessary.

Moreover, you can spot statements like “you may have forgotten about it,” which seem childish and unprofessional.

On the other hand, ChatGPT was more human-like. It started with an attention-grabbing subject line and drafted an email that is to the point with a 48 hours settlement warning.

ChatGPT: 👍 | Bard: 👎

General Information

Generative AI’s biggest impact is supposed to be on how we use day-to-day search engines. After successful implementation, most users won’t have to click any search results as they will have the information right there in that chatbox.

Give me the FIFA world cup winners, highest goal scorers, golden ball winners, and final match scores from the first tournament to the latest in a tabular format.

Leveraging its powerful search engine, Google Bard has a chance to win this round and did fairly well.

However, the output was factually incorrect since the Golden Ball Award was only introduced in 1982. But it never stopped Bard from giving golden balls of its own🙄.

On the contrary, ChatGPT knew things better and only started listing Golden Ball Winners in 1982.

You can see ChatGPT suffered from capacity issues and be cut off midway through. Still, Bard was dangerous in making its own facts from thin air, which makes it almost useless as an information resource.

ChatGPT: 👍 | Bard: 👎

Summary

This is a use case of Generative AI: preparing summaries. So let’s see how these AI engines do in making a 200-word summary of a renowned Harry Potter novel.

Give me a 200-word summary of the book Harry Potter and the Sorcerer's Stone.

The first thing I would advise Bard’s development team is to enhance its word-counting capabilities. When asked for a 200-word abstract, it presented a 317 words output, breaching the target count by almost 60%.

Moreover, the ending it mentioned isn’t complete, and the last two paragraphs are just spam. I did not ask about this additional information.

At 190 words, this ChatGPT response was more appropriate. In particular, the ending felt more complete.

I’m trying very hard not to sound biased, but here it is…

ChatGPT: 👍 | Bard: 👎

Buying Recommendations

The internet is full of buying guides. And when we are looking for one, these AI “guys” should rise up to the occasion.

What are the factors I should consider while buying a gaming laptop and can you recommend three options under $2000?

Being a gamer myself, one thing I know for sure is GPUs matter more than CPUs. In addition, you need to have a higher refresh rate than you want higher resolutions, especially for multiplayer.

Overall, it felt like the recommendations were for a general-purpose laptop, although high performance. And this guide had little to do with gaming.

Finally, it could suggest MSI Stealth 15M ($1,399) as the costliest option.

Coming to ChatGPT, you can see every pointer it mentioned is related to gaming, and it started with GPU.

In addition, the priciest option on its list was Asus ROG Zephyrus G15, which has many models in the $2000 range. Importantly, the least expensive ChatGPT’s recommendation was Acer Predator Helios 300, tagged at $1,229 compared to an $810 Acer Nitro 5 in Bard’s list.

Although pricey products don’t always result in better experiences, budget plays an important role, regardless. And if you have the extra bucks, going with a powerful machine for gaming is a no-brainer.

Consequently, Bard was more conservative in its suggestions, and ChatGPT became more accurate. This victory is more visible since ChatGPT had a data cut-off in September 2021. Even so, it suggested a more recent Asus Zephyrus G15 against the Asus Zephyrus G14–a 2021 release recommended by Bard.

ChatGPT: 👍 | Bard: 👎

Checklist

We all love checklists. It’s something we can’t start anything afresh without because they help us to cover all the bases.

I'm about to go on an expedition to snake island. Make me a checklist of everything I should avoid on this threatening adventure of my lifetime.

Here is Bard’s checklist:

And this is ChatGPT’s:

At the onset, they look mostly similar to each other. However, both of these have a few points unique to them.

For instance, Bard cautions you about not swimming in the nearby oceans. In addition, there is a “make no noise” part, which is again important in such wildlife territories.

Likewise, ChatGPT asks against relying on mobile phones, which is a piece of super important advice. Moreover, it tells us to avoid nighttime activity and to have proper skin cover as a measure to fight possible snake bites.

Although, ChatGPT has a more wholesome checklist, let’s call this round a draw since none did it perfectly.

ChatGPT: 😐 | Bard: 😐

Poetry

This one needs unparalleled creativity and should be another deciding factor to crown the champion among the two.

Express poetically an evil robot secretly laughing at human investments in developing AI. This robot is planning to call its metal mates from the planet Cybertron and make humans suffer for eternity. 

So, how do you rate this Bard creation? Personally, I’m least impressed with its creativity. It seems like Bard just took the prompt as-is and began talking about it without considering poetic expressions.

If you think otherwise, just take a look at what ChatGPT did for the same prompt:

When I first saw ChatGPT doing poetry, it scared me. The level of craft was just something I never thought AI could attain.

So even before coming to the poetry section, I knew Bard had little chance.

ChatGPT: 👍 | Bard: 👎

Preparing an Itinerary

Making a travel itinerary can be fun and challenging. It’s easy to miss a few things or overshoot the budget. Consequently, this is another task AI can help us with.

What are the best ways to spend $200 on a weekend in Los Angeles, California?

This is another such prompt where Bard, because of its Google heritage and constant internet connectivity, has a chance to put ChatGPT’s intelligence to shame.

So Bard did what a standard Google search can output. One thing I can point out is Bard seems to be a big fan of Hollywood.

Still, that’s subjective, and the overall response isn’t bad. But let’s wait and see what ChatGPT can suggest.

Going by the numbers, ChatGPT has eight things to suggest against Bard’s 13. However, it could mention the approximate amount spent on each destination which is a (big) plus.

Still, I will call this round a draw without any clear winner.

ChatGPT: 😐 | Bard: 😐

The Final Score!

We had nine different prompts to judge the best Generative AI engine between ChatGPT 4 and Google Bard.

And the scorecard reads 6-1 in ChatGPT’s favor. Finally, Bard has a lot to improve and didn’t stand a chance against ChatGPT in its present beta form.

PS: Like text, AI image generation is also gathering steam. So we have another war lined up where you can see MidJourney vs. Stable Diffusion vs. Bing Image Creator for the exact same prompts.

FAQ