Google and AI search summaries
Will AI-generated search summaries be good for the internet? Viewpoints from multiple sides.
New to Framechange? Sign up for free to see multiple sides in your inbox.
Learn more about our mission to reduce polarization here.
What’s happening
Last week, Google Search announced the rollout of the AI Overviews feature for US users with plans to roll it out to over 1B users globally by the end of the year. With Google Search’s estimated 4.9B users and 92% share of the search market worldwide, its expanding use of AI could have a significant impact on how people find information online.
What are AI Overviews: AI Overviews, which build on the Search Generative Experience (SGE) Google has been testing over the past year, essentially summarize an answer to a user’s search query, combining multiple sources with embedded links to more information. Answers are provided in narrative form directly underneath the search bar. The feature is designed to ultimately address complex searches such as “find the best yoga or pilates studios in Boston and show me details on their intro offers, and walking time from Beacon Hill.”
Broader AI search trend: Google’s launch of AI Overviews comes as generative AI companies make a bigger push into search. Perplexity AI, founded in 2022, similarly returns narrative summaries on search topics with annotated links to its sources. The company is valued at $1B+ and has 10M+ monthly active users.
OpenAI, which recently launched its most advanced AI model GPT-4o, is rumored to be developing a Google Search competitor. Arc Search is another AI-first search engine that launched this year.
Implications: Content publishers rely on traffic to their websites to generate revenue through ads or subscriptions. The emergence of AI search summaries has raised concerns among some publishers that AI search will reduce traffic to their websites when users get a satisfactory answer from the AI summary without needing to click through to the source.
Many users and tech companies, meanwhile, praise the efficiency and information access they expect to come with the emerging features.
This week, we bring you viewpoints from multiple sides on the implications of AI search summaries for the internet.
Notable viewpoints
Emphasizing the drawbacks of AI search summaries:
AI search summaries will reduce web traffic to content publishers and hurt their business.
Search result summaries that use content from a publisher will reduce the incentive for users to click through to the publisher’s website when most of the information they need appears directly in the AI-generated response.
AI search summaries that appear at the top of the search page will reduce the likelihood that users click on other websites not included in the summary and that appear further down the page.
Research firm Gartner projects there will be a 25% reduction in traffic to websites from search engines by 2026; content creators, who largely depend on ads to monetize their business, may collectively see $2B in losses from AI-generated search summaries within one year due to lower traffic.
“Google’s new product will further diminish the limited traffic publishers rely on to invest in journalists, uncover and report on critical issues, and to fuel these AI summaries in the first place…It is offensive and potentially unlawful to accept this fate from a dominant monopoly that makes up the rules as they go.” (Danielle Coffey, chief executive of the News/Media Alliance.)
AI search summaries will weaken the quality of information available on the internet.
If AI-generated search summaries reduce traffic to publisher websites, publishers will have less incentive to create content, which would result in fewer content creators, original reporting, and how-to guides and a general dearth of quality information on the internet.
Without compensation for use of their work or traffic to their websites, publishers will have less incentive to allow search engines to crawl their websites which could result in the fuel source for AI search summaries drying up, diminishing their usefulness.
AI-generated search results need to improve in quality and accuracy to be trustworthy.
Reports have found that many Google SGE summaries included made up facts, misinterpreted questions, and poor source quality.
The use of publisher content in AI-generated search results without compensation should not be legal.
The use of publisher content in AI training data and AI-driven search results should not be permitted without compensation to publishers; there have been multiple lawsuits filed by publishers against generative AI companies such as OpenAI, Microsoft, and Meta.
Many AI models still regurgitate facts and words directly from source websites; for example, there have been numerous examples of OpenAI answer results including text passages verbatim from the New York Times, which should be considered copyright violations.
Emphasizing the benefits of AI search summaries:
AI search summaries will dramatically improve the experience for users and accelerate access to information.
AI search summaries make the search experience faster and more informative.
Google’s AI Overviews provide contextually relevant answers to complex user searches and, through testing, have improved user search efficiency and satisfaction. (Summarized from Google blog post announcing AI Overviews.)
AI search summaries will help users find specific products they are looking for that may have otherwise been buried under pages of search results that traditionally prioritize products with strong search engine optimization (SEO) practices regardless of whether they are the best product match.
Google has an obligation to help its users find answers to questions they want answered as efficiently as possible.
Google’s AI Overviews have gotten more powerful and faster in a relatively short time period, and are compelling some users to use Google’s search engine instead of chatbots like ChatGPT.
Google AI Overviews will increase click-through rate and engagement with publisher content.
“[W]e see that the links included in AI Overviews get more clicks than if the page had appeared as a traditional web listing for that query. As we expand this experience, we’ll continue to focus on sending valuable traffic to publishers and creators.” (Liz Reid, Head of Google Search.)
In testing, users visit a greater diversity of websites for complex research questions when using Google AI Overviews than they do when using traditional search results. (Summarized from Google blog post announcing AI Overviews.)
The use of publisher content in AI-generated search results should be considered legal.
The use of publisher content in generative AI and AI-driven search results should be legal under “fair use” because it remixes and uses the content in transformative ways consistent with copyright law, rather than just verbatim copying.
As long as AI does not generate results that are substantially similar in wording to the source material, there should be no copyright violation.
Other viewpoints:
AI-generated search results will reshape search engine optimization (SEO) strategies for publishers that have grown used to gaming the system with copy that typically helps elevate their content in traditional search engines.
AI-driven search could drive the creation of millions of fake products by people trying to game the system and fulfill every possible combination of features that someone might search for.
Google’s AI Overviews feature is the latest in a long line of search features – such as Knowledge Panels and accelerated mobile pages (AMP) – that have steadily reduced direct returns for content publishers; for example, Wikipedia began seeing significant decreases in web traffic – a 14% decrease between 2019 and 2020 – after Google started featuring its content directly in search results.
Support Framechange by sharing it
Help us reduce polarization by spreading the word on Framechange. We believe our world would be a better place if everyone had a straightforward understanding of multiple sides. Click the button below to share with friends, family members, or colleagues. We'd greatly appreciate it!
Be heard
We want to hear from you! Reply to this email with your perspective on AI-generated search summaries and we may feature it in our socials or future newsletters. Below are topic ideas to consider.
Do you think AI search summaries will be net-positive or net-negative for the web?
What are some arguments or supporting points you appreciate about a viewpoint you disagree with?
Give us your feedback! Please let us know how we can improve.
#BTW
Listen to this moving, building Radiohead cover by indie artist Lianne La Havas.