After the latest updates to both OpenAI’s ChatGPT and Google’s Bard, it’s time to find out what are the big differences between both viral chatbots.
After ChatGPT was released to the public in November 2022, it was only a matter of time before other tech giants created their own versions. In March 2023, Bard, Google’s answer to the groundbreaking OpenAI chatbot, launched in the US and UK — though it has now been scaled but not to Europe.
These apps have been in the crosshairs of AI and other big companies for months as text-to-image, text-to-video and even text-to-music platforms have been developed.
At a basic level, both chatbots use natural language processing, which means that users enter a message or query, and the chatbots generate a human-like response. However, there are some key differences that come down to the data sources and the models they have been trained on.
In this article the great differences and uses of ChatGPT and Bard are collected so that you can get an idea of the advantages and disadvantages of each one.
These are the big differences between ChatGPT and Bard
Language models with which they have been trained
The key difference between ChatGPT and Google Bard can be found in the Large Language Models (LLM) they employ. OpenAI’s ChatGPT uses GPT-3.5 or GPT-4 in its paid version, while Google Bard is based on its custom language model for dialog applications (LaMDA).
At their core, the two models share the revolutionary architecture of the transformer, a neural network specifically designed to process sequential inputs.
GPT-3 and GPT-4 have trained on a wide range of text from the Internet. We are talking about websites, books, articles, documents, etc.
On the other hand, the Google LaMDA training was enriched with Infiniset, a dataset focused primarily on dialogues and conversations—Common Crawl, Wikipedia, published documents, and a wide variety of web-based conversations.
Of course, mentioning that relatively recently Google announced that Bard was now going to this under the umbrella of PaLM 2 —another of its powerful language models— which has allowed us to improve the initial responses, especially in terms of mathematics, reasoning and coding is concerned.
Doubts vs. text generation
On the other hand and having already analyzed the characteristics of both, Bard is designed for research, while ChatGPT is a great writer, so to speak. Bard is quite useful if you are looking for short answers on any topic as it finds the most relevant information on the internet and summarizes it for you.
ChatGPT, on the other hand, is more focused on generating text based on statistical patterns. As part of this tool, it works as a chatbot but can also serve to summarize texts, translate and other functions on a more textual level.
Internet vs. events up to 2021
Bard will pull information from the data and sources already available on the internet, whereas ChatGPT’s knowledge is limited to events up to 2021, Bard will have access to the most recent date and be able to provide more up-to-date information, which of course is quite a plus point.
This, in addition to being integrated into the Google search engine, will have access to a multitude of data, which will give it an advantage over ChatGPT, which is backed by Microsoft.
It is essential to know what the differences are between the two, but not with the aim of qualifying one as better than the other. As you can see, despite starting from the same base, their paths are oriented differently.
Perhaps what is striking about both chatbots is that they can be used in a complementary and not substitutive way. Perhaps therein lies the true potential of artificial intelligence and the use of these new tools.
