Bard: Google’s Answer To ChatGPT

According to Google’s CEO, Sundar Pichai, Bard is a new AI chatbot powered by their Language Model for Dialogue Applications (LaMDA). Two years ago, Google unveiled its next-generation language and conversation capabilities powered by LaMDA and has since been working on an experimental conversational AI service that will herald in a new era for Google and also help them deliver on their mission to “organize the world’s information and make it universally accessible and useful.” The time is today, and Bard is now officially available to the public. Of course, this raises some curious questions about what this means going forward for the AI industry and, as we will discuss here, its closest rival, ChatGPT. 

Bard, according to Sundar, “seeks to combine the breadth of the world’s knowledge with the power, intelligence, and creativity of our large language models. It draws on information from the web to provide fresh, high-quality responses. Bard can be an outlet for creativity and a launchpad for curiosity, helping you to explain new discoveries from NASA’s James Webb Space Telescope to a 9-year-old or learn more about the best strikers in football right now, and then get drills to build your skills” 

Google’s phrasing of what Bard does is by no means coincidental, and anyone familiar with ChatGPT would quite agree that the words “fresh”, “new”, and “right now” are not applicable to ChatGPT, whose knowledge base is limited to events that occurred on or before the year 2021. 

How does Bard work?

Bard works quite similarly to ChatGPT; upon receiving any query, it works to understand the query and then proceeds to search its database of information for the best possible answer. Perhaps what I find particularly interesting about Bard is that it’s connected to the internet in real-time. Because Bard is connected to the internet, it is constantly being updated with new information. This means that Bard is always up-to-date on the latest news, events, and research. This is a huge plus for Bard, as ChatGPT’s knowledge limitation is seen as a major disadvantage by users looking for the most up-to-date information.

Let’s take a closer look at some other benefits of using Bard. 

Export a response in two clicks.

Bard unlike ChatGPT allows you to export their response in either of these two ways;

1. Export to doc: By exporting to doc, you can directly create a new document straight out of Bard containing your generated response. Saving you the hassle of having to separately create a new document and then copy and paste into that new document. 
2. Draft in Gmail: Your generated response can be exported to a draft in your gmail with the click of a button after which you can then send to your desired recipient(s). I think this feature probably falls under the category of “features I never knew I needed.”

Voice Input

You can save a lot of time by speaking with Bard rather than typing. When you want to talk to Bard, you can do away with your keyboard and instead communicate with it through audio prompts. This is the closest you’ll get to feeling like Tony Stark conversing with Jarvis; it’s not perfect yet, so who knows what future updates will bring?

Access for Google Workspace accounts

If you’re a Google Workspace admin, then this is for you. You can now enable Bard for your domains, thereby allowing your users registered on your workspace to access Bard using their workspace account. This is a very significant addition, as it provides ease in engaging in research and helps with work and other business needs. 

Help with coding

While ChatGPT is limited to only 12 coding languages, Bard is said to now support over 20 programming languages such as C++, Go, Java, JavaScript, Python, TypeScript, and others. “When Bard generates Python code,” Google says, “you can also export and test the code directly in Google Colab.” Bard can be used by programmers of all skill levels, from beginners to seasoned engineers, to generate and export code, debug code, and explain how code works.

Broader range of interests

Google has added a ‘Google it’ button to Bard, which allows you to see more suggested search topics. So, if you enter a prompt and are unhappy with the generated response from Bard, you can use “Google it” to view search topics related to the prompt you entered.

Multiple generated response to your prompt

Whereas ChatGPT generates only one response to your prompt, Bard generates up to three different responses. These responses are labeled ‘Draft 1’, ‘Draft 2,’ and ‘Draft 3’. This almost certainly ensures that Bard generates something that corresponds to your prompt.

Overall, the most exciting benefit of a Google-backed conversational AI model like Bard is that, well, it’s Google-backed, which opens up a slew of potential integrations with many of Google’s other products. Even at this early stage, Google is attempting to create some sort of synergy between Bard and its other products, such as Gmail, Google Workspace, Google Docs, and Google Search. This is just the beginning of what’s to come, and we’ll be seeing a lot more Google products integrated with Bard in the near future.

Bard in action

I have had a lot of fun using Bard since it became public, and since it’s still under development, I wasn’t totally expecting the complete package, and in all honesty, I wasn’t disappointed. Bard for now appears to be wide off the mark with some of its responses, and I don’t quite doubt that some of the promises being made about Bard will either take a long time to perfect or not even work exactly the way they are being sold. Here are a few examples of the use cases for Bard;

I started out searching for the latest update on the 2023 NBA playoff, and I thought this shouldn’t really be a problem for Bard after all; it’s connected to the internet and it’s touted to provide “real-time updates”, not to mention it also has the full might of the Google search engine at its disposal. The result…

Hmmmmmmm.. what? All it takes is one quick search on the internet to write this off as incorrect, yet we see Bard presenting this as factual, which leaves a lot to be desired and also raises some important questions about the reliability of AI generative models like Bard and ChatGPT, who have a knack of creating completely fictional articles to back up their claims.

Not to be deterred by Bard failing its first and significant test, I tried using it to summarize a web page, and once again I was left a little disappointed with the result.

To say this is bad is an understatement, because not only did Bard incredibly fail to capture a chunk of useful information that formed the crux of the article, but it also invented completely new points that were not mentioned in the article while also focusing on the minor details that were barely mentioned in the article.

I also used Bard to create a sales pitch, and it’s possible that this is where I found some success.

Prompt: Act as a Sales person to convince clients to sign up for a free trial to Postly.ai

So far, so good. Bard is 1 for 3 in my assessment, and it’s only fair that we be a little forgiving by remembering that it’s still in its early phase and it’s not quite the finished product yet.

Conclusion

Going forward, I expect there to be a lot of updates as Google strives to perfect Bard and also integrate a lot more of their products with their AI model. This is a genuine reason to be excited about its potential to be game-changing and to also worry about the unprecedented changes this piece of technology will usher in and what the future holds.