How AI Could Disrupt Google
AI-powered chatbots such as ChatGPT could well disrupt Google’s business model within the next few years, says Paul Buchheit, a computer engineer, and entrepreneur who created Gmail and developed the original prototype of Google AdSense.
ChatGPT is the latest evolution of OpenAI’s GPT-3, a language model that uses deep learning to produce rather impressive human-like text. And unlike Meta’s Galactica which was pulled after three days or its ill-fated BlenderBot 3 in August, the consensus is that it is “scarily good”.
A year or two from ‘total disruption’
“Google may be only a year or two away from total disruption. AI will eliminate the Search Engine Result Page, which is where they make most of their money. Even if they catch up on AI, they can't fully deploy it without destroying the most valuable part of their business!” wrote Buchheit in a tweet.
But how is an AI-powered chatbot related to a search engine? In a series of tweets, Buchheit alluded to how Google killed a pre-Internet business – the Yellow Pages, and how AI would now do the same to the search engine.
“The Yellow Pages used to be a great business, but then Google got so good that everyone stopped using the yellow pages. AI will do the same thing to web search,” he wrote.
In his view, search engines of the future will be used by AI agents to gather relevant information and links, which are then summarized to give users the precise information they need without having to parse through multiple pages of search results.
Buchheit argues that this would be akin to a professional human researcher doing the work, with the difference being the ability of the AI to respond instantly.
Echoing the sentiment, Puplett, a tech founder, wrote in a Medium blog post: “As [large language models] become more advanced and widely adopted, they are likely to become the dominant form of question-and-answer (Q&A) technology, potentially even using voice input and output. This could have a profound impact on the way we access information and seek answers to our questions.”
Further improvements to AI ahead
Finally, the disruption will further exacerbate the widespread problem of plagiarism that search engines caused in schools. But now, the problem will become real-time.
As noted by science website IFLScience, Jon Agar, a professor at UCL recounted how he gave a seminar on Turing’s “Computing machinery and intelligence” last week. A student was asked to respond to one of Turing’s objections, and he came up with a “perfectly good answer”.
It turned out that the student had punched the query into ChatGPT and read the answer aloud.
“Although the student revealed it later, I did NOT notice anything wrong with the initial response. It was a good, intelligent seminar answer. This can now be done live and routinely in university seminar discussions,” wrote the professor.
Finally, while ChatGPT could probably be reasonably dubbed as “GPT-3.5”, the next major upgrade in the form of GPT-4 is expected to be released as soon as next year. Expect event more challenges ahead.
Paul Mah is the editor of DSAITrends. A former system administrator, programmer, and IT lecturer, he enjoys writing both code and prose. You can reach him at [email protected].
Image credit: iStockphoto/David Arment
