Challenging Google

New AI-driven search can be disruptive

Artificial intelligence (AI) programs involving Natural Language Processing (NLP) have caused a shift in the way people use web search. A battle between AI-driven search engines has begun, which may lead to radical changes in the way search is monetised via advertising, and it could conceivably challenge Google’s dominant search engine off its pedestal. Since its public launch in November 2022, ChatGPT has caught on at extraordinary speed. The AI crossed the threshold of 100 million active users within two months — this is way quicker than any previous program. The OpenAI laboratory, which created ChatGPT, has entered a partnership with Microsoft. Microsoft’s search engine Bing has incorporated ChatGPT and other AI technologies from OpenAI. Google has responded by unveiling Bard, its own NLP program. Every major browser is now looking at creating extensions that embed such tools. Apart from Microsoft’s Edge, Mozilla Firefox, Opera, and Chrome now allow searches to be presented in conventional fashion as well as in ChatGPT-style.

ChatGPT and Bard make behavioural sense because they offer more “human” filtering of search results than the conventional engine’s listing of relevant links. This is by presenting search results in an essay format, which explains the material, rather than just displaying links. Many users say this makes it easier for them to understand the content they are searching for. However, even as search results are presented plausibly and comprehensibly, this does not mean they are necessarily accurate. There can be over-reach in assertions. Conspiracy theories and opinions can be presented as facts if they are stated by “authorities”. Fact checking can also go awry. At its launch presentation for instance, Bard asserted the James Webb Space Telescope was the first telescope to take pictures of an exoplanet, which is wrong. Ironing out glaring defects like this could take many iterations.
 

ChatGPT and Bard (which is based on Google’s LAMDA engine or Language Model for Dialogue Applications) are also capable of nuanced, high-calibre translations, which is another big attraction. Their ability to write essays, or even poetry (in ChatGPT’s case) on demand, is another crowd pleaser though this may open the door to a new kind of plagiarism. It is likely that high-school and college students will use these tools to compose homework, and academics will have to watch out for that. AI-assisted fiction is another possibility that may change the publishing industry. This is quite apart from other applications, such as asking ChatGPT to write software code to perform various functions, or perhaps using NLP to build an autonomous driving application.
 

There are many use-cases for NLP that go way beyond web-search. However, what may be of immediate concern for investors and industry-watchers is that NLP-based searches could lead to an entirely new revenue model for one of the Web’s biggest market segments. It is possible some upstart could design new search tools around AI and grab market share from established players. That, in turn, would mean advertising revenue share moving to new players, or perhaps some entirely different revenue model. Since around 2004, Google has been dominating search and it has built an empire around the cash-cow of resulting advertising revenues. This is finally being challenged and any competition that challenges an established pecking order is usually a good thing.