The CEO of Google, Sundar Pichai, has announced that the company is planning to revamp its core search engine by incorporating a chat-based AI model, which will allow users to ask questions to Google and engage with large language models (LLMs) within the context of search.
While a specific timeline has not been disclosed, Pichai has indicated that the new feature will be rolled out in the near future, potentially within months.
According to Pichai, the opportunity for the new experience is greater than before, and it will offer a more user-friendly interface for search engine users.
The move follows the launch of Google’s AI chatbot, Bard, which was made available to a limited number of beta testers in March.
Currently, Bard is a standalone website that operates separately from Google’s core search engine, although it includes links to the mothership for factchecking and supplementary information.
Pichai has stated that Google is working to refine the product and prepare its infrastructure for a future public launch.
However, given the substantial computing power required by large language models like Bard, ChatGPT, and Bing, transitioning to the new system will be a costly undertaking, particularly for a company of Google’s size, which commands 85% of worldwide search volume, compared to Bing’s 9%.
Pichai has revealed that Google’s two AI teams, Google Brain and DeepMind, will work together to reduce the load on the system.
He expects to see greater collaboration between the two teams, particularly for more compute-intensive efforts. Google has also announced its plans to cut 12,000 jobs, representing 6% of its workforce, in the coming year.
When comparing Bard, ChatGPT, and Bing, one of the main differences is the way users can access them. Bing is currently the closest model to what Google may launch, as it is already integrated into the public search engine.
Upon accessing the page, users are prompted to engage with the chatbot first before being presented with a more traditional, link-based experience.
We reached out to Bard to inquire about how Google plans to integrate it into its flagship search engine, and the response was that Google has not yet announced specific plans for this integration.
However, there are a few possibilities being considered, all aimed at replacing a list of links with a concise summary of the information found within those links.
Regardless of the user experience, concerns around the ethical implications of AI’s potential to spread misinformation are likely to arise, particularly given that Google is the world’s largest search engine.
Recently, Elon Musk, Steve Wozniak, and other tech giants published a letter calling for a six-month pause on AI systems more powerful than GPT-4, developed by ChatGPT’s parent company OpenAI. However, AI ethicists argue that such a pause may not be the best approach.
It appears that a pause on AI development is unlikely, and instead, Sundar Pichai, CEO of Google, suggests that a more widespread integration of AI is on the horizon.
Pichai predicts that smaller AI models will become increasingly democratized and useful for companies and individuals to design and run their own algorithms on personal devices. He believes that there will be a diverse range of options available, and the technology will be more accessible than expected.