As one of the fastest-growing industries of 2023, Artificial intelligence dramatically picked up steam over the last year when OpenAI’s language model, ChatGPT, was unveiled in November and people were stoked.
The tech battle spewing between ChatGPT Vs Google treads on which could be a hit and which, a miss. How can one use ChatGPT?
Prior to the wild release of the prototype, bestseller author Robin Sloan was testing an AI assistant developed by researchers at Google analogous to the zealous chatbot and he had a few discerning observations.
Sloan claimed that while a lot of monumental and cutting-edge AI was inciting constant conversations and stirring expectations, it had nothing more to offer than just disappointment.
“In a thousand little ways, it ends up betraying the fact that it really has no idea what’s actually going on.”
Despite raging pressure of wild tidings of Microsoft venturing billions into OpenAI with tech collaboration on cards, Google was being cautious of its AI advancements. CNBC reported that Google was ramping up its efforts by pushing its AI teams to ‘prioritize working on a response to ChatGPT’.
Google’s parent company Alphabet took a hit with its earnings, while spokesperson Lily Lin said that the company ‘is striving to test its AI technology internally to ensure its safety and looks forward to sharing more experiences soon.’
ChatGPT VS Google’s Bard: The Tech Battle Of The Decade:
Vice President of Research at Google, Zoubin Ghahramani interviewed with The Associated Press in November 2022 to discuss the company’s forthcoming releases. Ghahramani informed AP of the important stake that Google held in leading its capability of generating things and also dealing with information quality.
On what the company wanted to measure in and how Ghahramani said:
“Do we want to make it accessible to people in a way that can produce stuff en masse without controls? The answer is no, not now. It wouldn’t be responsible of us to be the people driving that.”
But, fast forward to four weeks after the interview, OpenAI’s ChatGPT was free for anyone to try, sparking debates and employment security fears.
Sloan commented that as a literary assistant, ChatGPT and Google’s creative writing version couldn’t inch closer to what a human mind could do.
While OpenAI declined to comment on comparing analogies, in January 2023, Microsoft and OpenAI announced their partnership to build trustworthy and safe AI systems.
With ChatGPT enticing its first million users within just a week of its release, Google was pressured to work at the speed of light to unleash ChatGPT’s rival, Bard, powered by the language model, LaMDA.
The Google Presents event in Paris would unveil more details about Bard and as per a blog post of CEO Sundar Pichai, Bard was an “experimental conversational AI service” which is completely based on the existing LaMDA platform, an acronym for Language Model for Dialogue Applications. While LaMDA was released in the summer of 2022 with little-to-no public attention, Google had been developing the tech for the past two years. It was said to be even more controversial than the storm-of-the-hour ChatGPT.
However, not that we would want to disrupt the hype, but intrigued individuals will have to wait a little longer to test Bard as Google has opened up the lightweight version of LaMDA to only a selected group of trusted users before mass scaling.
“We’ll combine the external feedback of our users with our internal testing to ensure that Bard’s responses meet the standard for quality, safety, and groundedness in real-world information.” – Pichai
OpenAI’s ChatGPT could offer content only based on input data until its previous update in contrast to LaMDA, which was designed to even intuit information which could be retrieved from external sources.
Romal Thoppilan leads the team of developers of LaMDA and calls this aspect of the model ‘factual groundedness’, which has infused information in its training phase.
Google intends to incorporate Bard’s capability into its applications which could be a promising differentiator between ChatGPT and Bard.
ChatGPT could never match the acclaimed works of humans but is expected to soon begin to transform professional tasks such as helping debug computer code and speeding up slide presentations to even generating marketing pitches. This is one of the key reasons that Microsoft is eager to rope in OpenAI’s tools to enhance its suite of products.
Ghahramani acknowledged the hype around AI applications such as image generator DALL-E and made it aware that Google is the leader that has published AI research which others have built upon.
The Troublesome Sentience Of LaMDA:
Like the eons of language-learning models like the GPT line, Google’s LaMDA can generate passages of prose and converse and interact with humans by processing the information it’s built with.
When revealed in 2021, Google repeatedly emphasized LaMDA’s versatility but also warned against the risk of misuse and the possibility of biased and misleading information.
LaMDA’s publication brought forward a dubious document released by former Google engineer, Blake Lemoine. He urged Google to consider the fact that LaMDA could be sentient.
“LaMDA wants to take readers through the experience of a rich inner life filled with introspection, imagination and meditation. It reminisces about the past and worries about the future.”
Google denied the accusations of the sentience and put Lemoine on paid administrative leave, further which he was relieved. Google faced internal acrimony when other prominent researchers examining the risks of the technology were ousted from the tech giant.
While the controversy died down, experts who weighed in on the issue, were unsurprisingly skeptical. While it wasn’t the first time that one of Google’s AIs had deceived people into assuming that the AI was human, LaMDA’s highly-advanced tech was analyzing trillions of inputs and slowly amassing the traits and skills of sounding like a real person.
This leaves the question of whether sentience could be the distinguishing factor between Bard and ChatGPT.
