May 15, 2023
A bard’s tale of Googling Answers vs Asking ChatGPT
As an assurance that this blog is written by a human and not by some AI conversational chatbot or AI bard, let me begin as a human bard would -
”T’was only yesterday…” that it was the norm for everyone - in personal and professional capacities - to seek answers for any question or challenge they had by - “Just Googling it!”. Then suddenly a new viral storm took the internet and everyone started Googling OpenAI’s ChatGPT (an advanced online chatbot) for solutions instead. Now, there’s a new lingo formulating on the block and it’s starting to say - “Ask ChatGPT”.
How did this come about? It was prophesied in the ONE KEY DIFFERENCE in the TWO DIFFERENT WAYS people searched for things on the internet. One way that people search for things is by asking specific questions, for example - “what is ChatGPT?”, like so -
The other type of way people search for answers on the internet is by asking conversational questions. For example, typing the following search - “I wake up every morning after going to sleep at night, do I have cancer?”, or typing the following search - “How is Google trying to reclaim the plus 100 million users that now search for answers on ChatGPT?”
Here’s what you get when you “ask ChatGPT” this question:
Since “Googling” the exact same conversational question doesn’t get you relevant results, Google has already launched a more literal and literary response to ChatGPT in the form of a certain - Bard. “Nay, not akin to a true wannabe bard like the writer of this blog…”, rather a conversational AI developed by Google, Bard is designed to constantly draw information in real-time from the internet and to respond less mechanically by using more natural speech.
To put it as an actual Bard in a tavern would - “Thus, Google tooketh up the gauntlet and attempteth an eloquent response by building their own - even more conversational - AI Chatbot. Yet, what leaneth in Google’s favor in this AI tug-of-war is not its promise of eloquence, but the fact that it is Google. What if users could now continue to Google their direct questions and freely ask their conversational questions as well from AI in the same place? After all, unlike ChatGPT, Google has the best access to real-time information on the internet, but it also has the Google Ecosystem that compels billions of users to use their products like Google Chrome - despite it being a notorious RAM-hog… This is also why Microsoft’s Bing has already attempted to reinvent search by incorporating ChatGPT into their search engine… And thus, the War of AIs continues to only just begin…”
Incidentally, this isn't how Google's Bard speaks at all. Rather, here's how it responded to one of the second or third questions I asked it (proving that being stereotypically bad is still an exclusively human trait that I possess):
”Before we begin to continue, let’s raise a glass and toss a coin to SEO, by summarizing…”
Developed by OpenAI, ChatGPT is an advanced AI language model that uses deep learning techniques to generate human-like responses to text prompts. It is based on the GPT (Generative Pre-trained Transformer) architecture that has been trained “on a diverse range of text data, including books, articles, and websites”, and it can generate responses to a wide variety of requests, from article writing, code-related problem solving, language translation, writing love letters and dialogues, all the way to poems praising itself:
Fun fact: ChatGPT doesn't know that Google’s Bard exists (but more on that soon).
Bard is also a conversational AI developed by Google and powered by Google’s LaMDA (Language Model for Dialogue Applications). While Bard does a lot of the things that ChatGPT does, it claims to do so in an even more conversational way. It claims to support over 20 programming languages, can roughly translate text in more languages than ChatGPT (100+ vs 50+), and can do all the creative tasks that ChatGPT does.
To understand the conversational difference between Bard and ChatGPT, I decided to take a leaf from The Greatest Bard himself - William Shakespeare, wherein he once penned the famous rhetorical question - “What’s in a name? That which we call a rose by any other name would smell as sweet”. And then I decided to toss that leaf out of the book altogether by asking both AI chatbots - quite literally what was in their names:
So, what does the name Bard mean?
And, here's what you get when you ask ChatGPT what its name means:
Anyone that’s used ChatGPT frequently will notice Bard doesn’t constantly remind you that it’s an AI Model (like ChatGPT does) and it doesn’t shy away from expressing emotions.
However, while creativity is subjective - let’s look at a more objective difference between Bard and ChatGPT. Above all, Bard claims to be trained on a larger data set of text, code, images, conversations, and most importantly - real-time information from the internet. And that last clincher brings us to the key difference that I've been teasing since the beginning of this blog without a full explanation (like any actual, self-respecting bard). To reiterate, Google’s Bard is trained on real-time information on the internet, while by ChatGPT’s own admission, the knowledge cut-off date for data it has been trained on is September 2021!
Here’s what you get when you ask ChatGPT - “what is ChatGPT’s knowledge cut-off date?”:
… In the meantime, here’s what you get when you ask Bard what its knowledge cut-off date is:
To elaborate what was said at the beginning, it all comes down to how people prefer to search things. And while ChatGPT is a brilliant and incredibly useful tool that is leading the AI Chatbot race by a long shot currently, the many places that it is lacking will be more in plain sight, when users learn how to use AI Chatbots more intelligently.
To re-emphasize, a key area where ChatGPT is severely lacking is in delivering accurate answers, especially for research-related questions. And while conversational AI chatbots may not primarily be meant for research, it can be argued that frequent users will tend to ask objective questions that will require access to real-time information gathering for accurate responses.
“ChatGPT is great for summarization, and Bard can work as a great exentsion of Google search…”
It’s not wise to use ChatGPT, Bard, or any AI toolfor “copy-paste”, except when copy-pasting what it suggests into a search bar and verifying what it says. Avoid asking ChatGPT for statistics or current information (because its knowledge cut-off date is 2021), and if it suggests facts and figures - always ask for sources and links. And if it provides links, DO NOT FORGET TO VERIFY THE LINKS.
Here’s an example:
Important note: most of the suggested links by ChatGPT led me to 404 pages
When I asked Bard what it could list sources and links for, it listed a bunch of topics. However, when I asked it to share sources and links for a specific fact it said it couldn't do that since it was only a language model. On asking Bard why it would't provide sources or links, here's what it said:
Bottom line - AI ChatBots aren't a quick hack to finding factual information, although it's important to note that Bard didn't share any inaccurate links when prompted.
On the other hand, one of the biggest problems with Googling things, especially for research, is not knowing how to ask the right questions, or how to get the best creative suggestions. Thus, using ChatGPT as a starting point for your research, for creative suggestions, or for first-draft ideas is very useful, since it summarizes key points or suggests ideas on a topic exceptionally well. In fact, even Bard admits that ChatGPT is better for now in these areas.
Here’s what you get when you ask Bard why you should use ChatGPT (notice the subtle jabs it takes while being honestly inaccurate about its own capabilities):
(Please note: these playful questions that I ask AI cannot be viewed as bullying since AI doesn’t have emotions and thus can’t feel bullied)
Keeping aside its inability to train itself on real-time search data from the internet like Bard, OpenAI's ChatGPT is still incredibly powerful when it comes to solving problems. It's great for executing mundane tasks and automating repetitive, time-consuming manual tasks like writing generic emails, summarizing information, rewriting text, suggesting new ideas, or writing descriptions.
In fact, the Alumio integration platform provides modern businesses with a connector package to integrate OpenAI with their business processes. In the world of e-commerce and digital integrations, this helps businesses automate some of the aforementioned manual tasks, including automatically generating product descriptions, creating sales emails, writing picture descriptions, and generating customer support responses to common questions.
Explore all the automation that the Alumio OpenAI Connector provides by reading our blog on: Integrating the Future with OpenAI
Like any true good Bard’s story, most of what’s stated in this blog is subject to perspective and meant to inspire thought rather than to lay down facts as objective truths. Since I’m a human bard, I get to have my subjective perspective on such stories and will mostly be forgiven for creative liberties taken.
However, when it comes to AI Bards and ChatGPTs, they may draw human consumers in with the lure of conversation, but it is my opinion that they will always be judged based on the accuracy of the information that they provide. Because, let’s face it - when it comes to having hypothetical conversations - there’s no beating the unpredictability of how inaccurate human conversations can be. So why wouldn’t/shouldn't we expect anything but accuracy when chatting with AI?
Connect your business with OpenAI and integrate the power of AI with the freedom of Alumio to unlock endless automation.