What to Know About How Google Is Reimagining Its Search Engine With Generative AI

It will help identify synthetically generated content through watermarks

Introducing the Adweek Podcast Network. Access infinite inspiration in your pocket on everything from career advice and creativity to metaverse marketing and more. Browse all podcasts.

To keep up with its dominance in the search market, Google is bringing generative artificial intelligence features to its Search engine to enhance search results in the coming weeks, the company announced at its I/O developer conference.

When a person types in a query into Google’s main search bar, they will see an AI-generated response pop up. Instead of just links to websites, the AI-driven model will write answers in full sentences, similar to ChatGPT, as well as surface links to sites with information backing up the response. This is in addition to the traditional search results. People will also be able to follow up with the queries through pop-up question prompts via a chatbot format. It’s unclear when the features will roll out and who will be able to access them.

“We have been applying AI to make our products radically more helpful for a while,” Sundar Pichai, CEO of Google, addressed the crowd at the conference. “And with generative AI, we are taking the next step. We are reimagining all our core products, including search.”

In its new effort called Search Generative Experience (SGE), Google will continue to allow search ads to appear in dedicated ad slots throughout the page. The ads will continue to feature Google’s transparent ad labels with the “Sponsored” label in bold black text.

In the coming weeks, SGE will roll out an experiment in Google’s Search Labs to test how, when, and in what formats ads show up. The tech giant will also gather advertisers’ feedback on their experience.

Google captures 92.6% of the global search engine market, according to April stats from Global Stat Counter. Meanwhile, Microsoft’s Bing accounts for 2.8% of the global desktop search market. In February, rival Microsoft introduced AI to its search engine Bing, spurring more people to sign up. Earlier this week, Microsoft outlined to advertisers how formats would work in AI-driven Bing.


“With Microsoft now consistently integrating AI-powered features into search, they’ve put pressure on Google to do the same,” said Monica Ho, CMO at martech company SOCi. “Brands must accelerate their efforts to improve their presence on search platforms by using AI and keep up with consumer expectations. Additionally, brands must ensure their data is kept up to date on search platforms and need to collaborate with their tech partners to ensure that information is incorporated into AI search results.”

Combating misinformation and ownership queries

Risks of AI-based content spurring misinformation have loomed, reaching world leaders. Earlier this month, the White House hosted a meeting with the CEOs of Google, Microsoft, Anthropic and OpenAI to discuss safe and responsible ways of developing AI.  

Google announced several tools to identify synthetically generated content, including watermarking and metadata capabilities, which are expected to roll out in the next few months.  

Google plans to incorporate metadata and markup in the original file of its AI-generated images so people would be to see labels marking images as AI-generated. Creators and publishers will be able to add similar metadata. The tech giant also plans to watermark its latest generative models to help with the challenge of misinformation.

“Metadata allows content creators to associate an additional context with original files giving you more information whenever you encounter an image,” said Pichai.

Although a step in the right direction, Chirag Shah, professor in the Information School at the University of Washington, points out that this attempt to solve for misinformation is limited to just images.

“If Google or anyone is pretending that this will entirely solve misinformation, that’s not true,” he said. Further, bad actors can still hack and manipulate images. And most end users may be limited in knowledge to check metadata or even question whether the images were generated using an AI tool, Shah added.

Google will also add two new ways for people to evaluate images. First, on Google search, people will be able to see key information, such as when and where similar images first appeared and where else the image has been seen online, including news fact-checking and social sites.

In addition to changes in its search engine, Google is scrapping the waitlist for its existing chatbot Bard—which recently made headlines for generating misinformation—and opening it up to people in 120 countries and 40 languages.

The tech giant also announced over 25 products and features, including coding capabilities, powered by PaLM 2, Google’s own AI language model and a rival to OpenAI’s GPT-4.

Enjoying Adweek's Content? Register for More Access!