The AI Arena: Evaluating Bard & Bing's Prowess

The AI Arena: Evaluating Bard & Bing’s Prowess

Google Bard was released shortly after Microsoft’s Bing Chat, and both tools are hoping to grab a healthy share of the burgeoning generative AI and integrated AI search market. But what are the differences between them? And which chatbot is right for you?

MUO VIDEO OF THE DAY

SCROLL TO CONTINUE WITH CONTENT

Disclaimer: This post includes affiliate links

If you click on a link and make a purchase, I may receive a commission at no extra cost to you.

What Are the Differences between Google Bard and Bing Chat?

Outwardly, both tools look similar. They both offer a simple interface with a text box to allow you to “chat” with them. However, dig a little deeper, and significant differences become apparent. To test the tools for yourself, you will need a Google and Microsoft account, although you might have to join a waitlist for Google Bard .

To compare Google Bard and Bing AI Chat, we derived a simple test. We used them to conduct research for an imaginary article on atmospheric CO2 levels and climate change. This enabled us to draw comparisons on several metrics, including the quality and size of the response, the accuracy of the data, and ease of use.

Bing Chat vs. Google Bard: Response Quality

We began by asking both platforms an identical question: “How do modern-day CO2 levels compare to the last ice age?”

There were differences between how both models approached the question. First up, we will look at Google Bard’s answer.

As you can see, the answer seems comprehensive and details current and historic CO2 levels in parts per million (PPM). By contrast, Bing Chat took a different approach.

The first notable difference is the approach taken in quantifying CO2 levels. Rather than list CO2 levels in terms of ppm, Bing Chat presented the data in terms of percentages.

The length and tone of the responses also differed. The Bing Chat response was shorter and more restrained, not offering opinions on the problem or potential solutions.

Bard produced a longer answer that also discussed the problem and the need to take urgent action. This was outside the focus of the question, but it confirms just how quickly the tone and direction of any research can be swayed depending on the tool you use.

Bing Chat vs. Google Bard: Response Accuracy

Before we look at this, it is important to note that both tools make clear that the accuracy of the information generated cannot be guaranteed. As such, we are not going to check every detail. Rather, we will look at how easy it is to fact-check the response of each tool.

Based on the answers to our original question, one major difference was apparent. Although the Bing Chat reply was less specific with the information, it did supply links to the cited sources. On the other hand, Google Bard provided no links in its response.

Furthermore, Bing Chat also allows you to scroll down to the “traditional” Bing search engine. Bard has included a Google it button, which activates a hyperlink that launches Google in a new tab. These are both useful features that allow you to check the supplied info against further sources.

This is probably good because one other interesting feature in Bard’s reply was the ability to view other drafts. Although the figures in the other drafts were similar, there was variation.

The main takeaway from this section is the importance of checking the accuracy of the information received from any chatbot.

Bing Chat vs. Google Bard: Ease of Use

There is nothing complicated about either model. If you have used a messaging app, you can use these tools. However, we did notice some differences that are relevant here.

We liked the fact that Bing Chat allowed you to scroll down to the Bing search engine. The Google it button was similar but involved opening new tabs. It’s not a big issue, but the Bing option was just that bit easier to use.

Perhaps more telling was another issue we found when testing the tools. We found ourselves puzzled by Bard’s response to a simple request to “write some bullet points for a blog on global warming.” This should be a request that any chatbot should handle. But, as you can see, Bard seemed confused by the request.

Whereas Bing Chat handled the request perfectly and responded with some bullet points.

We did manage to get a response from Bard by rewording the prompt.

As both platforms are going through rapid development, the user experience will continue to improve.

What Are the Differences Between the Language Models?

The last major difference regards the “engines” that power both platforms. These Large Language Models (LLMs) are the underpinning technology. We don’t need to go into great detail, but LLMs are important when choosing a platform.

Bing Chat uses GPT-4 to generate results, whereas Google’s LaMDA powers Google Bard. As a result, the architecture of the models is different, and this can affect the answers.

GPT-4 is a generative model. It generates answers based on the text it’s been trained on. By contrast, LaMDA works differently —it is a discriminative model, meaning it can distinguish between various kinds of text.

In essence, this should mean that GPT-4 is better at generating creative content, while LaMDA is better at answering questions about facts.

Which Is Best for You: Bing Chat or Google Bard?

This is a case of choosing the right tool for the right job. When we compare the performance of both platforms, Bing Chat seems to edge it. However, there are some provisos to consider. The first is that the Microsoft offering is a more complete product, while Bard is still rough around the edges.

But don’t rule Bard out. We liked the more fact-driven approach that Google Bard took; the inclusion of ppm levels of CO2 was closer to the spirit of the original question than the Bing response. This is what we would expect to see, given its underlying architecture.

Another useful feature of Google Bard was its ability to return multi-draft options on a single prompt. This was lacking with Bing Chat.

Finally, the platforms that the tools run on are also a consideration. For example, Bing Chat is restricted to running on the Microsoft Edge Browser, while Google Bard will run on any supported browser.

Both tools are incredibly powerful and can generate intelligent answers to any prompt. But both have pros and cons, and the ultimate choice is likely to be decided by factors like user preferences and the desired output type.

They’re free to use, so we advise trying them both. These are impressive tools that are changing the way we use the internet.

The Era of the Chatbot Is Upon Us

Regardless of the flaws we found and our quibbles with the chatbots, these are amazing technological feats that will play an ever-increasing part in our lives. The rapid development of technology will continue to push the boundaries of what these tools can achieve and how we use them.

As AI-powered chatbots like Google Bard and Bing Chat become more sophisticated and capable of handling complex tasks, they will become an even more integral part of our daily lives.