The introduction of ChatGPT shook several tech firms. Particularly concerned is Google, whose revenue is mainly reliant on its search business, which ChatGPT may someday jeopardize.
Less than three months after ChatGPT’s introduction, Google has announced the creation of Bard, an AI designed to challenge ChatGPT’s apparent hegemony. But how does Bard operate? Will Bard surpass ChatGPT in quality? Here is what we currently know.
What Is Bard AI from Google?
Bard is a Google-owned experimental AI language model that may produce responses that sound human in response to commands or inputs. It is constructed on top of LaMDA, Google’s potent conversational language model (Language Model for Dialogue Applications).
“Sundar Pichai, the CEO of Google, revealed Bard on Google’s The Keyword blog, giving a face to a long-rumored product that Google had in development to compete with ChatGPT.”
Unlike ChatGPT, Bard will initially only be accessible to a small number of trusted testers, according to Sundar Pichai, with intentions to scale to more users. Additionally, the announcement provided various samples of the initial release, some of which are eerily similar to what ChatGPT will achieve. Will Bard only be another ChatGPT, then?
Why Google Bard Might Outperform ChatGPT
On November 30, 2022, ChatGPT was made public following a protracted development phase. The fundamental technology that underpins it was developed over the course of several years. This would appear to suggest that Google had less than three months to develop a product to compete with ChatGPT, a product based on years of developing technology.
The truth, though, is quite different. For many years, Google has made significant investments in artificial intelligence sectors including natural language processing. In a strange turn of events, Google is attempting to compete with ChatGPT while also developing the Transformer architecture, the technical architecture that underpins ChatGPT. To put it another way, Google is currently attempting to defeat a danger to its business that is technically based on its own technology.
Sundar Pichai was keen to highlight ChatGPT’s groundbreaking status in creating Transformer models in its Bard launch blog post, seemingly in reference to ChatGPT. He emphasized that Google research serves as “the foundation for many of the generative AI applications” that are now in use.
Beyond his words, the proof is unambiguous, nevertheless. Because of this, GPT-3, which powers ChatGPT, and LaMDA, which powers Bard AI, are extremely similar technologies. So no, Google is not a fresh player in the competition to develop a multifunctional AI chatbot. In contrast, the business has been developing comparable, if not superior, technology for years and may have a major advantage over OpenAI when it comes to who eventually takes the top spot.
That’s not all, though. Additionally, Google benefits from having more data at its disposal. More data may lead to improved outcomes in the AI industry, particularly when it comes to training conversational models like GPT-3 and LaMDA. It’s unclear whether and how Google plans to incorporate real-time data or online information into Bard’s responses. If this does occur, it would, nevertheless, be a substantial improvement over ChatGPT’s pre-trained method of reacting to requests.
Simply put, this would mean that Google’s Bard could offer current, relevant responses, whereas ChatGPT would be restricted to information about events, not going past 2021 (ChatGPT’s current data training cut-off).
However, despite the excitement and logical anticipation around Google’s Bard announcement, its initial presentation failed to impress. Astronomers were eager to point out that at least one of Bard’s “facts” was wholly incorrect. For instance, Bard claimed that the European Very Large Telescope “got the very photos of a planet outside of our own solar system,” when that feat was actually carried out by the James Webb Space Telescope in 2004.
Google has been hoarding a variety of very extraordinary AI technologies while only sometimes revealing them to the public.
With OpenAI’s audacious wager to expose ChatGPT to the public, with all the danger it entails, Google may finally overcome its concern about reputational harm and showcase its capabilities. If the perks of Google’s AI models over the years are any indication, Google Bard may actually be a serious rival to ChatGPT