There is an epic battle brewing between two titans of technology, and its outcome will radically alter how we use the internet forever. In an increasingly heated AI-race, Microsoft and Google are now duking it over the future of search, which is poised to become one of the most disruptive developments in decades.
What is ChatGPT?
You may have heard the internet and social media buzzing about ‘ChatGPT‘ recently, but is it exactly? Put simply, it’s a chatbot, built by the San Francisco-based OpenAI, designed to mimic human conversation. The party trick here, and what allows it to so closely approximate real, human communication, is that it was trained and refined on large language models. Hence the official name, Chat Generative Pre-trained Transformer, although that’s admittedly not quite as catchy.
The technology powering ChatGPT is based on a form of artificial intelligence, in that the extremely large amount of data it was trained on allows it to now make predictions on what it should write, with a high degree of accuracy. This training occured through reinforcement techniques along with supervised back-and-forth conversations with humans, to refine the model and improve its dialogue format.
In its current form, the chatbot works by providing lifelike responses to human queries, going beyond simple question-response or data-lookup behaviour of similar systems you may have interacted with in the past. Unlike those, ChatGPT can handle a wide variety of questions and provide an equally impressive array of answers. For instance, it can help you write an essay, debug computer code, compose music, write poetry or even predictively answer questions in the same tone of voice as your favourite celebrity. And the results are stunning, to say the least.
How ChatGPT Set Off Alarms At Google
A chatbot that can deliver fast and accurate answers, in a human-friendly format (instead of a page of endless links) is cause for concern for Google, the dominant player in search. However, the tech giant doesn’t appear to be taking this development lightly, as the arrival of ChatGPT seems to have sent shockwaves through the company. Amid announcing it was laying off more than 12,000 employees and refocusing its efforts on AI, the company has sought advice from co-founders Larry Page and Sergey Brin on how to respond to this challenge. The pair were summoned to emergency meetings, after previously stepping down from their co-CEO roles in 2019.
Make no mistake, the successful public demonstration of AI-powered search is an existential threat for Google.
Make no mistake, the successful public demonstration of AI-powered search is an existential threat for Google. After all, revenue from search and advertising is more than just a strong segment, it is the company’s lifeblood. The most recent earnings report paints a picture as to just how dependent the company (and parent company Alphabet) is on search, especially as revenues continue to slump each quarter (the sixth consecutive). Q4 2022 saw search bring in $42.6 billion USD, while Google made $76.05 billion overall – a sign of how little the company is diversified. That’s why the search giant simply can’t afford to lose this market.
Let The AI Wars Begin
Google isn’t the only one eyeing this new technology. In fact, an old rival has been hard at work behind the scenes. In 2019, Microsoft invested $1 billion into OpenAI, following that up with another $10 billion dollars supplied earlier this year. The ongoing financial commitment has led to a close collaboration between the two companies, with Microsoft benefiting from the technologies being developed by OpenAI on the former’s Azure cloud supercomputer infrastructure.
“AI will fundamentally change every software category, starting with the largest category of all – search.”
Satya Nadella
This week, Microsoft saw the fruit of their investment, with the announcement that ChatGPT would be coming to Bing, the company’s fledgling search engine. Powered by an upgraded version of the AI technology behind ChatGPT, the ‘new Bing’ is available for users to try out right now, albeit in a limited preview. There are several demo-prompts available to choose from, including asking the search engine to provide help with writing a story, guidance on coding or suggesting art project ideas for your child. You’ll then be taken to a regular search experience, with what looks like standard results, save for the addition of a side-panel displaying additional responses, supplied by AI. You can also join a waitlist to be given access to the full experience that will be coming at a later date.
The company is also promising to infuse AI into other its software offerings, including the Edge browser, Office and more. The company is dubbing this effort ‘your copilot for the web’, leaning into the idea that AI is a tool to help you do more by helping you cut down on tedious tasks and manual processes.
By adding AI-enhanced features to the Edge browser alongside AI-powered search, it’s hoped that users will be able to access information faster and with more relevance than traditional methods. Speaking to the company’s plans, Microsoft Chairman and CEO, Satya Nadella, noted that “AI will fundamentally change every software category, starting with the largest category of all – search.”
In The Race To Perfect Search, We All Win
The announcement by Microsoft came just one day after Google’s CEO Sundar Pichai delivered the news that they too were working on adding AI to search. Known as Bard, the company’s ChatGPT competitor was announced in a brief blog post on February 6th. Pichai promised that Bard would be “more widely available to the public in the coming weeks”, with the post providing some detail on the underlying technology. This rushed announcement, most likely triggered by the now accurate rumours that Microsoft was set to reveal its own AI-search product, is a clear sign that the company is worried.
Whilst AI and the underlying technology that powers ChatGPT is not novel, inviting the public to play with these new toys on such a large scale is something that has never been done before. This has led to widespread discussion about how AI will impact the future of work, education, and of particular concern to Google, how we will search the web. The company has traditionally remained cautious of AI, and there are plenty of public examples of AI-gone-wrong that somewhat justify their hesitation.
Unfortunately, the first public demonstration of Bard didn’t go to plan. During a demo of its AI-search integration, the Google bot proclaimed that the James Webb Space Telescope “took the very first pictures of a planet outside of our own solar system.” Only it didn’t, as that honour goes to Chauvin et al. (2004) back in 2004, as cited by Reuters and confirmed in this post by NASA. The fact that Google itself didn’t fact-check its own AI application ahead of the demonstration speaks to the rushed nature of its response. Now, it’s early days, but thus far Microsoft has been able to avoid these blunders, in part due to more carefully curating the experience during its own special event this week where it presented the new technology to select news and media outlets. Either Microsoft’s presentation was more carefully controlled, or its claims of powerful AI-smarts really are justified.
But is all this another example of rushing to adopt new technology that were aren’t yet ready for? Not quite. Both OpenAI, Google and Microsoft claim to put a focus on delivering results in a responsible manner, by adhering to AI principles around misinformation, screening for harmful inputs and seeking to mitigate novel risks through user feedback.
And of course, the information provided is only as good as the source, which is why it is reassuring that Microsoft is including references in its results – something that the original ChatGPT doesn’t offer. This could address concerns around plagiarism and non-factual information, as the new Bing will invite users to do their homework and check the source of its search results.
AI-assistive technology affords endless opportunities for users to save time and enjoy information in a more accessible and relevant manner.
Competition is what’s best for consumers, and these latest developments provide an early glimpse into some wonderful possibilities. AI and conversational search could yield massive improvements for accessibility, allowing those with cognitive or physical impairments a more natural and easier to navigate interface by which to search for information. Instead of relaying links and snippets of meta-data, screen readers and text-to-speech tools could plausibly take on a human-like tone of voice, acting more akin to a helpful assistant than a piece of technology attempting to synthesise speech from text that was foremost designed to be read. Likewise, those who are more responsive to auditory learning will benefit from the conversational dialogue, an option that was previously missing from the search experience.
Ultimately, this new leap forward in AI-assistive technology affords endless opportunities for users to save time and enjoy information in a more accessible and relevant manner. The next generation of search is here, and it will be exciting to see how it evolves over the coming months and years ahead. If this week is anything to go by, we won’t have to wait too long for the next exciting development.
■
