ChatGPT vs. Bard: What’s the difference? [2023]

We’ve all witnessed the hype around ChatGPT since OpenAI released it to the public. Chances are you’re one of its 100 million users—I know I am. I’ve used the AI chatbot to generate content ideas for articles, write emails I’d rather not write, and get Google Sheets formulas I can’t figure out on my own. (Which is often.) 

While I’ve perfected the art of writing prompts to get the best result, I am painfully aware of ChatGPT’s limitations: for example, in order to access the web for more accurate answers, you have to pay for ChatGPT Plus, and even then, the web browser can be a little slow.

This is a key difference that Bard, Google’s answer to ChatGPT, definitely scores points with—it offers a lot of what ChatGPT does, faster and for free.

Since the release of Bard, I’ve been using both tools; one as a research tool, the other as a writing assistant. But there are plenty of other differences that set them apart—and which will dictate how you use each tool. Let’s dive in.

ChatGPT vs. Google Bard at a glance

At a base level, both chatbots use natural language processing, which means users key in a prompt or query, and the chatbots generate a human-like response. 

There are a few key differences, though, that boil down to the data sources and models they’ve been trained on.

• Google Bard uses Google’s Pathways Language Model (PaLM 2) and can offer responses based on real-time, current events pulled from Google Search—making for a great research tool. 

• ChatGPT, on the other hand, uses its Generative Pre-training Transformer 3 (GPT-3), or GPT-4 for paid subscribers, and while it can draw responses from Bing search, it makes for a better text generator or writing tool. 

I’ll walk through some of the core differences between ChatGPT and Google Bard in depth in the coming sections, but here’s a quick breakdown of how they compare.

Bard offers a better user experience

Bard is lightyears ahead of ChatGPT in terms of its user-friendly interface. Not only does it just look nicer—with formatted text that’s way easier to scan than ChatGPT’s chunky text—but you can also edit your questions after you ask them and view multiple responses that it prepares.

And every response has a CTA that says Google it, so you can confirm its sources. Plus, it’s connected to your Google Workspace, making it easy for you to upload Bard’s responses to your Gmail or to Google Docs.

The main downside to Bard’s usability is that it doesn’t really store or keep track of your previous conversations. 

ChatGPT, in contrast, lets you see everything you’ve ever asked it in the side panel. You can even share those conversations with friends and co-workers, making it a much more collaborative experience.

When the person you shared the link with opens it up, your conversation will be transferred over to their ChatGPT interface, making it easy for them to see your hilarious conversations, detailed prompts, or whatever else it is you’re sharing—and it’ll let them pick up where you left off with the memories the conversation stored.

Bard is built for research, whereas ChatGPT is the better writer

Let’s start with this: both Bard and ChatGPT will run into logic issues. For example, I asked them both the same simple logic question, and they both got it wrong.

GPT-4 (using ChatGPT Plus) got it right, but it still messes up plenty of other things. Bottom line: it’s pretty hard to compare the accuracy of the outputs—especially since a lot of it depends on how you phrase the prompt.

But lack of reasoning skills aside, Bard’s output is pretty helpful if you’re looking for succinct answers around any topic. That’s because Bard finds the most relevant information across Google and summarizes it for you. The benefit here is that you don’t have to click through different pages or compare information, as Bard will do that for you.  

The result? Bard’s one very chatty assistant, zealous in its approach to gathering research for you to make your life easier—citing sources as it goes. 

GPT, on the other hand, is the industry standard when it comes to natural language tasks, powering other AI tools like Jasper, Copy.ai, and Bing’s AI tools. It’s highly trained on web text and more focused on generating text based on statistical patterns. As part of ChatGPT, it functions as a chatbot, but it can also serve as a summarizer, a translator, and other roles on a more textual level. 

This makes GPT (in my opinion) a much better writing tool than Bard. As a quick example, when I asked GPT-4 to write a tweet about Zapier, it did so pretty much in line with the rules for writing tweets (e.g., short character count, emojis, hashtags).

It’s not amazing, and I’d definitely kill every exclamation point and tweak the copy, but it’s a working draft. Bard’s copy, on the other hand, feels a little stale.

As for the bullet points it lists, it’s almost like it wants me to figure out how to write a tweet. Which, fine, that’s great for someone who’s learning the basics, but maybe less great if you just want a working draft right away. Of course, a little prompt engineering might change things, but this gives you a sense of how the two tools operate differently.

I found that ChatGPT is also better at brainstorming blog ideas, writing long-form articles or emails, and coming up with content marketing ideas. When I asked ChatGPT to provide me with an outline for a blog post about the crisis surrounding bees, the depth of detail in its output far exceeded Bard’s.

So, while GPT can understand and generate a wide range of text for multiple purposes (including content marketing), Bard feels like it was designed primarily to act as a research tool.

Bard draws from Google Search, whereas ChatGPT’s web browser is powered by Bing

Obviously, Bard would be a pretty shoddy research tool if it couldn’t tap into the internet. Unlike ChatGPT’s free version (which is basically a giant encyclopedia that cuts off in 2021), Bard will naturally scour Google for answers its training data set isn’t privy to. 

This is particularly relevant when it comes to:

• Researching recent events

• Summarizing key information in web pages

• Surfacing images for visual context

Recent events

I asked Bard about recent discoveries made by the James Webb telescope, and it instantly parsed these results:

As you can see, the way it distills that information is pretty neat. For one, it gives me trustworthy links, so at least I know the results are accurate. It also describes each discovery with an easy-to-grasp explanation. 

ChatGPT Plus (the only way you can access the internet using ChatGPT) is a lot less user-friendly. For starters, you have to remember to click on the web browser, which is powered by Microsoft Bing. 

When asked the same question, ChatGPT faltered a bit. For starters, it took it a while to generate a response. I was left waiting on this for a solid 30 seconds.

When it finally stopped scouring the web, it gave me two discoveries (Bard gave me four), and I couldn’t exactly scan its response quickly for the takeaway points, because it’s…chunky. 

The source was ok (CNN), but I do wonder why it didn’t favor publications that are primarily focused on space and technology. At one point, it told me it failed to click on a link (something that won’t happen in Bard, because it’s not clicking links in the same way), so it could just be hopping onto the next best thing. But as a whole, you’re better off just Googling your answer. (Or using this method.)

This is important, though: I asked Bard for discoveries in the last six months—and it gave me year-old results. ChatGPT, on the other hand, actually found a discovery as recent as the current month. So for all of its sluggishness in getting into gear, it actually gave me a more up-to-date response. 

This was somewhat curtailed by the fact that it hallucinated a date of its next discovery (I’m writing this in June 2023), but other than that, the details were completely accurate. 

Speaking of hallucinations, Bard’s also prone to muddling its responses. When I asked about a book that was recently published by Kate Morton, this was its response:

It gets the name of the book, the link, and the initial plot right. So far so good. But when you scroll down for the longer synopsis, that’s where things get weird. While it may sound convincing, the characters mentioned are completely made up, and it’s not set in World War II. 

My guess is that Bard is better for research that has plenty of web content to sift through, whereas, for something like new books that are behind a paywall, it tends to hallucinate. Probably because it doesn’t have enough material for an educated response. 

At least ChatGPT plays it safe.

This all serves as a reminder that both tools are far from perfect—and both require constant fact-checking. 

Summarizing web information 

Any research tool must be good at summarizing content or articles, particularly if the subject matter is complex. This is where Bard currently excels.

All you need to do is drop in a link to an article (this one’s about water guns for adults), and ask it for a summary:

Amazingly, it actually breaks the contents down into bullet points with accurate information. I double-checked in case it hallucinated, but it was spot on. 

This is pretty big stuff: for any research paper, medical journal, or webpage with a complex subject matter—you can ask Bard to give you a detailed recap, then even ask follow-up questions if there’s something you don’t understand. You can see how this might really change the nature of research itself. 

Water guns don’t really fall into the category of complex subject matter, but you get the gist. 

If you’re using ChatGPT, the only way to get it to summarize articles for you is if you copy and paste the text from the article into the prompt box. AndChatGPT has a limit of 4,096 tokens, which is equal to roughly 3,000 words. For anything longer (like research papers), you’d need ChatGPT Plus. 

Speaking of which, let’s take a look at how ChatGPT’s paid version handles summaries. Like Bard, you can just drop in the link to the article, like so: 

Again, it generated a chunky response, which might not be ideal if you’re looking for a quick overview. Bard knows automatically to give you a quick summary in the form of bullet points, ideal for scanning. 

However, you’ll see that ChatGPT was able to capture certain nuances Bard didn’t, like the fact that the toy market is shrinking due to birth rates dropping. This is pretty important if the beginning of a study dictates the reasons behind the rest of the topic, so it’s good to bear in mind that ChatGPT takes in the finer details. 

And remember: you can specify how long you want your summary to be in your prompt, so even if your first attempts are a little wordy, that shouldn’t stop you from using it as an effective summarizer. 

Surfacing images from the web for visual context

Bard’s ability to surface both search results and images from Google Search definitely sets it apart as a major player. Although it can’t create images using AI (yet), it can surface images that already exist according to the topic you’ve asked it about. 

Whether you’re researching specific dog breeds, Vermeer paintings, or even bicycle repairs, Bard can churn out specific images from other web pages for visual context. 

It also includes the source it grabbed the image from, so you can click through the links if you want to check the sources (or visit the actual web page they’re from for more information). 

Bard’s ability to surface visuals just gives you a better sense of what you’re exploring:  

It’s just another feature that marks its place as a research tool—aided by Google Search.

ChatGPT has more integrations and a whole suite of plugins

ChatGPT offers a massive range of plugins that unlock even more use cases, far more than Bard can offer.  

For example, with the Expedia plugin, I just tell ChatGPT about a trip I’m thinking of booking, and it’ll immediately surface the cheapest flights it can find via the travel site, along with the link, airport details, duration of flight, and most importantly, the pricing.  

It’s so much easier than going through travel sites yourself, adjusting filters, and comparing sites side-by-side. It will also offer lodgings or other area-specific activities for you to explore. 

You can ask Bard to do the same, but it will provide you with completely made-up prices. And in this case, the images it provides (logos of the airlines) are just distracting.

You can even install a number of ChatGPT plugins to work in tandem. For example, you can ask the AI for a recipe recommendation, get an accurate count of calories (using the Wolfram plugin), and then ask it to create a shopping list (with the Instacart plugin). 

It’s practically like having a personal assistant—ideal for those who hate planning ahead for anything. (Ahem.) 

You can also build ChatGPT directly into your existing workflows with Zapier’s ChatGPT integration. That means you can use ChatGPT to write content and generate replies across your favorite apps. Here are a couple of examples.

Bard vs. ChatGPT: Which is better?

That was a lot to take in, so let’s do a quick rundown of the pros and cons.

Google Bard: pros and cons

ChatGPT: pros and cons

The better AI tool depends on what you’re using it for—and whether you can come to terms with those pesky hallucinations. 

Related reading:

• How to use AI badly

• The best AI productivity tools

• How to use OpenAI’s GPT to write business emails

This article was originally published in March 2023. The most recent update was in June 2023.