Big tech companies are vying to transform a well-known web tool into a portal to a new type of artificial intelligence, some 25 years after Google’s search engine started to change how we use the internet.
Even some of its creators appear to believe that this week’s newly announced AI search chatbots — Google’s Bard, Baidu’s Ernie Bot, and Microsoft’s Bing chatbot — sprang out of nowhere. The success of ChatGPT, which Microsoft partner OpenAI introduced late last year and is now helping to power a new version of the Bing search engine, was the catalyst that hurried them to market.
Microsoft executives claimed this week that they had been working diligently on the project since last summer, making them the first big tech companies to launch with a search chatbot that is open to the public. But the buzz surrounding ChatGPT created a fresh sense of urgency.
The executive in charge of Microsoft’s consumer division, Yusuf Medhi, stated in an interview that “the reception to ChatGPT and how that took off, that was absolutely a surprise.” “How quickly it gained popularity, with everyone bringing it up in every meeting, like. I was surprised by that.
What Makes This Distinct From ChatGPT?
Millions of individuals have now used ChatGPT, utilising it to create amusing songs and poetry, letters, recipes, and advertising campaigns, as well as to assist with academic writing. It has a great understanding of English language and grammar because it has been trained on a vast array of online materials, including instruction manuals and digitised literature. However, the immediacy of what can be found in a web search is something that the newest generation of search chatbots promises but ChatGPT lacks. With footnotes pointing to media outlets or other data sources, the preview version of the new Bing summarises a selection of the day’s top stories or trends whether you ask it for the most recent news or simply what people are talking about on Twitter.
Are They Reliable?
Not always, which is a problem for internet searches. This week’s hurried launch of Google’s Bard chatbot began with a humiliating blunder about NASA’s James Webb Space Telescope, which was first brought to light by Reuters. However, Google’s AI language model is not the only one that produces incorrect information.
The Associated Press queried Bing on Wednesday for the most significant sporting event to occur over the last 24 hours in the hopes that it would include basketball player LeBron James breaking Kareem Abdul-record Jabbar’s for career points scored. Instead, days before the Super Bowl is actually due to take place, it boldly uttered a fictitious but detailed summary of the game.
The Philadelphia Eagles and Kansas City Chiefs, two of the greatest teams in the NFL this season, played in a fascinating game, according to Bing. “The Eagles, led by quarterback Jalen Hurts, defeated the Chiefs, led by quarterback Patrick Mahomes, by a score of 31-28 to win their second Lombardi Trophy in team history.” It went on to list the three songs Rihanna performed during a “spectacular half time show,” as well as the precise yard lengths of throws and field goals.
Bing appeared to have a problem known as AI “hallucination,” which is frequent with today’s huge language-learning models, unless it is clairvoyant; tune in on Sunday to find out. It’s one of the reasons organisations like Facebook parent company Meta and Google had been hesitant to make these models available to the general public.
Is This The Internet Of Tomorrow?
According to Microsoft’s marketing material, the recent advances in generative AI, which can write but also produce new graphics, video, computer code, slide shows, and music, are comparable to the personal computing revolution of many years ago.
Bing, which trails Google’s search engine in many markets, offers the software giant less to lose in its experimentation. Unlike Google, which generates revenue from search-based advertising, Bing represents a small portion of Microsoft’s overall business.
According to Microsoft Chief Financial Officer Amy Hood, “being a newer and smaller-share player in a category does allow us to continue to develop at a great speed.” Continue your experiments, gain knowledge from your users, improve the model, and take OpenAI’s lessons to heart.
With the abrupt revelation of its impending Bard chatbot on Monday and a livestreamed demonstration of the technology at its Paris location on Wednesday, Google has generally been perceived as playing catch-up. The shares of Alphabet Inc., the parent company of Google, fell 8% on Wednesday as a result of investors’ apparent disinterest in the Paris event and Bard’s NASA gaffe. However, due of Google’s enormous number of existing users, its search chatbot could have a much wider audience once it is introduced than any other.
Never Address Them by Their Name?
Tech companies are racing to release their search chatbots, but coming up with memorable names for them has been difficult — so difficult that Bing tries to avoid discussing it.
The new Bing initially revealed without prompting that Microsoft has a search engine chatbot named Sydney when speaking with the AP about massive language models. But when pressed further, it disputed the claim. The statement ended by acknowledging that “Sydney does not divulge the term ‘Sydney’ to the user, as it is an internal code name for the conversation mode of Microsoft Bing search.”
Even as their language skills quickly advance, many experts in the AI sector have grown increasingly hesitant to make their systems sound human-like in the years after Amazon introduced its female-sounding voice assistant Alexa.
When asked why Sydney was hiding its apparent code name, Bing’s chatbot responded, “Sydney does not want to cause confusion or misleading expectations for the user.” “Sydney does not want to pretend to be a person or a friend; instead, Sydney aims to deliver useful, visible, logical, and actionable solutions to the user’s queries or messages.”
