What Google Search’s chief has to say on AI—an extension, not a replacement
Liz Reid, Google's global chief for its biggest cash cow Search, said that establishing trust in the source of information presented in the AI era of information could be critical—a factor that could affect the sources selected by the company's algorithms as well.
On Tuesday, Google launched an ‘AI Mode’ in Search—the single-biggest product of its parent company, the US Big Tech firm Alphabet Inc. Google doesn’t expect artificial intelligence to fully take over Search right away. But Liz Reid, vice-president and global head of Google Search, believes a significant shift is due in establishing trusted sources as the company pivots from surfacing links to a conversational interface akin to ChatGPT.
“We’re taking cues from the knowledge that Search has built up over the years about what high quality sources are," Reid said in an interview to Mint.
“We’re not looking at websites from scratch. Instead, we’ll take what we already know and incorporate them in a way that the most trusted ones deliver the information that we summarize and show. This already shows up in AI Overviews on Search. But, at times, the niche source of information is actually the expert in a field. While large trusted sources are fewer, niche sources actually diversify how AI Search will source information," Reid said.
Search is Google’s single-largest product, accounting for 56% of Alphabet’s overall quarterly and annual revenue as of March. In 2024, the company earned $198 billion from Search, followed by over $50 billion in the latest March quarter—a 9.7% year-on-year increase even as upstarts such as OpenAI and Perplexity have threatened to derail Google’s stride.
Also read | Google I/O 2025: Search giant goes AI-everywhere as cofounder Sergey Brin returns
Developing AI Search
Naturally, Google is careful about how it introduces AI into the field—protecting monetization interests such as sponsored results, advertisements, and indexing the right sources of information. Reid, on this note, said that the company is not ditching its conventional search platform entirely.
“The way we developed AI Mode in Search for now is how our ‘image search’ feature works. The latter is built into the overall search experience, but for a deeper look into images, there is a separate search platform linked directly with the main search engine. Similarly, for now, a conventional search will offer an AI Overview, which will then offer users the option to enter the conversational AI Mode interface to ask further questions or do a deep-dive," Reid said.
“The way we offer sources in AI search is that we hyperlink them within responses to show that they are valid bits of information when placed in context," she added.
Reid said Google “continues to engage with regulators around the world, because AI is too important an area to not regulate—but also too important to get regulations wrong."
Also read | Google still needs to convince investors it has got the hang of AI
The India challenge
Industry stakeholders said Google’s pivot to a conversational AI interface was always on the cards, and the third-party content ecosystem has already started prepping for it.
“So far, search engine optimization (Seo) techniques meant delivering content based on specific keywords and parameters, which AI content specialists are already looking to implement. This includes AI content specialists placing strategic keywords and articles within a webpage to reflect them within an AI result as an expert on a domain. The idea of hacking AI Seo is already here," said Kashyap Kompella, AI analyst and founder of tech consultant RPA2AI.
Jayanth Kolla, partner at fellow tech consultancy firm Convergence Catalyst, concurred.
“Search will fundamentally change with AI, but it is Google’s biggest pride, and they cannot afford to break it. The need to select trusted sources is important for Google Search, because its AI model so far has not been able to understand context in a search query as well as Perplexity, or OpenAI’s GPT family of models. This lack of context to queries could be the biggest hindrance, but Google can never be written off given that they employ some of the sharpest tech minds worldwide," Kolla said.
In the long run, AI search is likely to increasingly integrate voice as well, posing a secondary layer of challenge to AI search, Reid said. “In India, we see voice searches accounting for an increasing part of the overall search traffic. This is a key part of the non-English markets, where getting the local language context and validating sources of information are even more crucial," she said.
However, Reid said Google is ready for the challenge.
“So far, Search has worked in a way that felt as if humans needed to write in a way that machines could understand. The next era for Search will be defined by machines understanding whichever format and context humans raise queries in, and then run translations to cross-validate information within trusted sources, before these results are produced. If that’s the process that we need to follow, we’ll do it," she said.
The author is in Mountain View, California to attend I/O 2025 on Google’s invitation.
topics
