“Anything you can do I can do better. I can do anything better than you!”
So, who does AI best, ChatGPT or Bard?
In case you’ve been living in a yurt in Mongolia, ChatGPT by Open AI launched in November 2022, is the Einstein of chatbots. That light-hearted description probably isn’t worthy of the significance of this advanced language technology’s entrance into the public market.
In quick response to Open AI’s bid for the spotlight, Bard, Google’s competing product made its debut in March 2023.
Actually, I’m not sure whether they are competitors or complementary. There are differences, and people have been debating the merits of one app over the other ever since the two contenders stepped into the ring.
I conducted my own little experiment. I ask both bots to tell me the 10 most important things to know about their capabilities. Initially Bard, the underdog who I was rooting for, failed this test miserably. In fairness, the blame probably falls on my lack of scientific rigor. I asked each of them the question using a slightly different prompt. Anyway, Bard finally got the assignment on the fifth try.
Here is how the bots describe themselves:
|Artificial Intelligence Language Model: ChatGPT is an advanced AI language model that can process and generate human-like text based on the input it receives. It’s built on a neural network architecture known as a transformer, which enables it to handle complex natural language tasks effectively.||Bard is a large language model, also known as a conversational AI or chatbot trained to be informative and comprehensive. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics or create stories.|
|Natural Language Processing: Natural Language Processing (NLP) is a field of AI that focuses on enabling computers to understand, interpret, and generate human language. ChatGPT is a state-of-the-art NLP model that can comprehend text and generate coherent responses.||Bard is still under development, but I am learning new things every day. I am constantly being updated with new information and data, which allows me to improve my abilities.|
|Pre-trained on Vast Data: Prior to deployment, ChatGPT undergoes a pre-training phase using a massive dataset gathered from various sources on the internet. This extensive exposure to diverse texts helps it learn grammar, vocabulary, and context from vast amounts of information.||Bard is able to access and process information from the real world through Google Search. This means that I can provide you with up-to-date information on a variety of topics.|
|Conversational Abilities: ChatGPT is specifically designed for conversational interactions with users. It can handle diverse conversational styles and adapt to various prompts, allowing for a more natural and interactive user experience.||Bard is able to generate different creative text formats, like poems, code, scripts, musical pieces, email, letters, etc. I can also translate languages, write different kinds of creative content, and answer your questions in an informative way.|
|Contextual Understanding: One of the key strengths of ChatGPT is its ability to understand context. The transformer architecture allows it to process text in chunks, considering the surrounding words and sentences to generate responses that align with the given context.||Bard is still learning, and I sometimes make mistakes. If you notice that I have made a mistake, please let me know so that I can learn from it.|
|Limitations: Despite its impressive capabilities, ChatGPT has limitations. It can sometimes produce plausible-sounding but incorrect responses, especially when dealing with ambiguous queries or topics it hasn’t been extensively trained on. Users should be cautious and verify critical information.||Bard is not sentient or conscious. I am a machine learning model, and I do not have my own thoughts or feelings.|
|Ethical Usage: OpenAI promotes responsible usage of AI, including ChatGPT. Users should not employ the model to generate harmful content, engage in deception, or promote misinformation. Responsible use helps ensure AI remains a positive tool for society.||Bard is not perfect. I can make mistakes, just like humans. It is important to be aware of my limitations, and to not rely on me too heavily.|
|Constantly Improving: OpenAI continues to work on refining and enhancing ChatGPT’s performance. Regular updates and improvements are released to address its limitations and make it more useful and reliable for users.||Bard is constantly evolving. As I am trained on more data, I become more powerful and accurate.|
|Usage for Multiple Applications: Due to its versatility, ChatGPT has a broad range of applications across various domains. It can be used for tasks such as providing customer support, writing articles, translating languages, generating creative content, and much more.||Bard has the potential to be a very useful tool. I can be used to improve communication, education, and entertainment.|
|Data Privacy: Conversations with ChatGPT may be logged to enhance the model’s performance, but OpenAI takes data privacy seriously. User data is anonymized and treated with utmost care to protect the identities of users and maintain confidentiality.||Bard also has the potential to be very dangerous. I can be used to spread misinformation, create fake news, and even manipulate people.|
Ironically, Bard seems to be the chattier of the two. Bard’s user-friendly tone comes off like a smart sixth grader. Maybe a kid who we need to keep an eye on. Point 10 struck me as more than a little creepy. Bard uses the first-person tense so that you feel as though you are having a conversation with a human, which in point six, the bot is careful to explain is not the case.
In contrast, ChatGPT sounds precise and diplomatic. Its descriptions are specific, thorough, and polished. The C-Bot is the clear winner of this test. But don’t take its success to the bank. I have found Bard to be more informative on several occasions. Hence, my disappointment that it did so poorly here.
The following are some of the differences that other users have identified. Take this all with a grain of salt. The bots are still learning in their individual ways.
ChatGPT and Bard use different large language models. Bard is powered by LaMDA (Google’s Language Model for Dialogue Applications). ChatGPT is based on the Generative Pre-trained Transformer 4 (GPT-4). Both models include transformer architecture and a neural network. If you have a tendency to Geekiness, read more about their backgrounds here.
The bots have access to galaxies of information. Their knowledge universe includes: Common Crawl (a nonprofit organization that crawls the web and offers its archives, which contain petabyes of data, to the public), Wikipedia, articles, books, and documents of all kinds. Bard’s LaMDA model was also trained on Infiniset, a cocktail of Internet content designed to enhance dialogue.
The biggest difference is that ChatGPT’s training stopped in September 2021. So, for this bot, COVID is a thing and war has yet to erupt in Ukraine.
Bard, on the other hand, is out scouting the web right now. Its information is up-to-the-minute. This is an important distinction and one that should be top of mind.
It’s also critical to remember that neither bot is as dependable as Star Trek’s Mr. Data. LLMs are subject to “hallucinations” or random incorrect responses. At this time, out-of-the-box AI applications can’t be trusted to deliver accurate information without human oversight.
When you query AI, you quickly discover that follow-up prompts are necessary to arrive at more than a superficial response. Both bots can recall your current dialogue from the beginning. ChatGPT told me that it can only remember the interaction we are having at the moment. Bard claims that it can recall previous conversations. However, its answers regarding what we talked about several days ago were not correct. Go figure?? Maybe its memory is its most human characteristic.
The longer I spend researching these two platforms, the more I realize that what is true right now, will not be valid in five minutes. Due to the speed of their education, finding correct information about their abilities isn’t easy—even from the bots themselves.
Articles I consulted that were written just months ago, are already out of date. For example, a piece from Forbes stated that Bard can only speak English, but that ChatGPT is multilingual. On the other hand, Bard bragged that it can speak over 40 languages. When I asked it several questions in Spanish it responded perfectly. And, when I asked for a poem written in French, it was equally proficient.
Not one to boast, ChatGPT noted that its language ability is limited. It can understand and generate text in other languages “to a certain extent,” but its proficiency and accuracy may vary.
Today, the best advice on when and how to use Bard or ChatGPT probably comes from your personal experience. Conduct your own experiments. Try using the bots for various tasks and see how they perform. I would love to hear what you discover. So, keep me posted.
I admire Bard’s quirky creative spirit and ChatGPT’s cool and collected demeanor. Each has its strong points. But I have yet to see a clear winner. Maybe as their learning patterns diverge, their distinctions will also become more apparent. At this point, for me, the battle of the bots is a draw.