Generative artificial intelligence (AI) has quickly risen in popularity over the last few years. It is incredibly effective at generating human-like text. While this technology is extremely versatile and has many applications, it is threatening the dominance of traditional search engines like Google Search.
Search engines are all about solving problems, and when you interact with one, it is a transactional relationship. Generative AI models like ChatGPT are quickly hijacking this relationship and solving transactional searches with refined results. And as these models are constantly improving and fine-tuning based on millions of feedback points, it is only going to get more difficult for search engines to keep up.
Traditional search engines use sophisticated algorithms to scan the web and provide users with relevant search results. However, generative AI is competing with these search engines by generating its own content that is just as informative, useful, and efficient as traditional search results. At the same time, it learns from vast amounts of online data, text, and user feedback on the internet, becoming more effective with each additional use.
Generative AI can also generate highly personalized responses that could prove more useful than the one-size-fits-all approach used by traditional search engines. It can create content that is more engaging and entertaining as well.
As users become more accustomed to conversational interfaces and messaging apps, many will begin to prefer engaging with these models over using search engines. This preference will become more pronounced as individuals learn how to engineer better and better prompts to get the results they were looking for. This will reduce Google’s dominance in search market share, which as of January 2023, is at 84.69 percent, and a decline in the popularity of Google Search as individuals shift towards conversational interfaces and messaging apps.
Google is not going to let its search engine be taken over by generative AI without a response. The company is investing heavily in AI and machine learning, and it is constantly improving its search algorithms to provide more accurate and relevant results.
Google’s parent company Alphabet is introducing an AI-powered chatbot named "Bard" to compete with Microsoft-backed ChatGPT. Bard uses Google’s Language Model for Dialogue Applications (LaMDA), which is trained on large amounts of dialogue to create natural conversations. The LaMDA used by Bard is a smaller version that requires less computing power, allowing more users to try it out.
Unlike ChatGPT, which is trained on data from 2021, Bard can retrieve information from the internet and provide fresh, high-quality responses. Users can ask Bard to explain complex topics or compare different items, such as the difficulty of learning piano versus guitar.
However, implementing AI search has proved to be a challenge for Google, as it requires more computing power and electricity than traditional searches.
According to Alphabet Chairman John Hennessy, AI search through an AI language model like Bard can cost the company up to 10 times more than a typical keyword search.
The cost of AI search varies, with some analysts estimating that Google could incur an extra $6 billion by 2024 if it answers half of its search queries with 50-word responses from an AI chatbot. Other analysts predict that integrating the chatbot into Google’s search engine could cost the company an additional $3 billion.
The reason why Google Search does not generate singular answers like ChatGPT has to do with the company’s business model, which is primarily focused on generating revenue through advertising. In 2021, 81% of Alphabet Inc.’s $257.6 billion revenue came from advertising, with much of that revenue generated by Google’s pay-per-click ads. This means that anything that could potentially prevent people from scanning search results and clicking on ads could significantly hurt Google’s transactional business model.
The shift towards generative AI will negatively impact Google’s advertising-based business model. Sridhar Ramaswamy, the former head of Google’s ads and commerce business, has pointed out that generative search from systems like ChatGPT will disrupt Google’s traditional search business "in a massive way." If users are presented with a single, accurate answer to their query, they may be less likely to click on the links provided by Google’s search results, which would hurt the company’s advertising revenue.
While Google is trying to address these problems by exploring new advertising models, such as "sponsored answers" that could potentially integrate advertising with AI-generated search results, they are behind in this race.
As generative AI technology continues to evolve, it will be interesting to see how Google and other search engine providers adapt to these changes and continue to offer users the best possible search experience.