The Emergence of the Post-Search Era on the Internet
By: Rooz Aliabadi
A year ago, I read an article discussing users’ mounting outrage and irritation with Google Search as automated summaries, sponsored content, advertising, and SEO-centric spam increasingly replaced the informative website results that the search engine was designed to produce. Rather than providing us with the information we were seeking (such as, in my case, the perfect toaster), Google’s search algorithm was inundating us with half-formed recommendations of “content farms.” However, Google Search has maintained its primacy due to habit and the absence of a viable alternative — until now. On February 7th, Microsoft initiated the beta rollout of an iteration of its Bing search engine as an A.I. chatbot powered by GPT-4, the most recent version of OpenAI’s large language model ChatGPT. Instead of directing us to external websites, the new version of Bing can generate answers to any inquiry. For a good reason, Google perceives this technology as an existential threat to its core enterprise. In late 2022, Microsoft issued a “code red.” Microsoft’s vice president of design, Liz Danzico, who contributed to developing Bing AI’s interface, recently said, "We’re in a post-search experience.”
The Bing A.I. combines Microsoft’s search directory and ChatGPT, which I recently tried. Using it is like conversing with an incredibly powerful librarian whose domain encompasses the vast expanse of the Internet. Nowadays, using keywords to search on Google has become second nature to most internet users like me. After entering the relevant keywords, we hit “enter” and peruse the list of links on the results page. They might return to the Google Search page and adjust their keywords if they don’t find what they want. However, with Bing A.I., websites act as source materials rather than destinations, and the bot collaborates with us to produce results. Bing A.I. filters through the information overload by summarizing the summaries and aggregating the aggregators. For example, I asked for Wirecutter’s recommended toaster, which provided me with the Cuisinart CPT-122 2-Slice Compact Plastic Toaster. I then asked it to gather a list of other suggestions, and it gathered them from various outlets, including Forbes, The Kitchen, and The Spruce Eats. Within seconds, I had a digest of reliable devices without leaving the Bing A.I. page. Nonetheless, the chatbot informed me it could not make my purchasing decision as it was not human.
A user of Bing A.I. has greater control than a Google Search user. We must learn to phrase their requests in complete sentences rather than isolated keywords when communicating with the chatbot. They can further refine their results by asking follow-up questions. For example, if we ask for an itinerary for a trip to Portugal and then ask, “What time does the sun set there?” the chatbot will understand which “there” we are referring to. However, in other ways, Bing A.I. limits us and encourages them to rely on the machine to determine helpful information rather than conducting their searches. The interface for Bing A.I.’s “conversation mode” is intended to be a one-stop shop for all our needs, from travel guides to financial advice. The interface consists of a single chat box on top of a subtle gradient of colors, and the chatbot even concludes its responses with a smiling, blushing emoji: “I’m always happy to chat with you. 😊” To the left of the chat box, there is a “new topic” button with a broom icon that clears the current conversation and starts over. The module was developed with the assistance of the A.I. itself.
Although Bing A.I. and similar tools may provide unprecedented convenience, they could harm content creators. While Bing A.I. does provide links to relevant websites, these are discreetly displayed as footnotes to minimize our effort. In a recent public video, Microsoft's Sarah Mody showed how Bing A.I. could reproduce an entire recipe within the chatbox, effectively circumventing the website that initially hosted the content. Mody then asked Bing A.I. to list the recipe’s ingredients and organize them by grocery-store aisle, a task that no recipe website could match. These features suggest that tools like Bing A.I. have the potential further to diminish the traffic and revenue of content creators.
Afterward, I requested Bing A.I. to provide me with the most recent news on the unfolding banking crisis, specifically First Republic Bank and SVB. Bing A.I. generated a summary of breaking news, citing articles from NBC, CNN, and the Wall Street Journal, which is behind a paywall. Although the Wall Street Journal has indicated that any A.I. that references its content must pay for a proper license, it may struggle to enforce this requirement for publicly accessible articles since A.I. search engines, like Google, crawl the entire Web. Then, I asked Bing to present the news in a bulleted list in style, a newsletter, and the result was a somewhat dry but convincing imitation. On another occasion, when I asked Bing for suitable wallpaper options for bathrooms with showers, it provided me with a bulleted list of manufacturers. Instead of searching for a listicle on Google, I “co-created” one with the bot.
The current design of the Web is heavily centered on aggregation, such as product recommendations on The Strategist, film reviews on Rotten Tomatoes, and restaurant reviews on Yelp. However, the rise of A.I. tools like Bing A.I. raises questions about the value of these sites in the future. Rather than relying on these sites for aggregation, we may bypass them entirely and rely solely on A.I. chat summaries. This paradoxically creates a reliance on the source material — the same information that other sites make — to generate answers. I believe the widespread adoption of A.I. tools could create a vicious cycle in which sites’ business models, based on advertising and subscriptions, collapse due to decreased direct traffic, leading to less content for A.I. tools to aggregate and summarize.
Regarding the potential impact of AI-generated content, Google and Microsoft recently introduced a suite of A.I. tools for office workers, including applications that can generate new emails, reports, and slide decks or summarize existing ones. These tools will likely extend into other areas of our digital lives as they become more ubiquitous. This could lead to “textual hyperinflation,” where it becomes difficult to distinguish between meaningful and meaningless content. A.I.-generated spam on an unprecedented scale could inundate us, and it may be challenging to differentiate between human content and machine-generated content. In such a scenario, “content mills” could use A.I. to create entire articles; publicists might write press releases using A.I., and cooking sites may use it to generate recipes. The glut of content may require human navigation assistance, but media companies may need more resources to devote to this need. However, A.I. may ultimately solve the problem it creates, as if tools like Bing A.I. cause the well of original material online to dry up; all that may remain are self-referential bots, offering generic answers that machines created in the first place.
As more and more content online is generated by artificial intelligence, I believe the non-automated text will become a sought-after commodity, akin to a natural and unprocessed product like natural wine. Google recently launched its own A.I. chatbot called Bard, which is a move in the ongoing competition between tech giants. However, Google has kept Bard separate from its flagship product, with one executive stating that it complements Google Search. This approach acknowledges the potential threat that A.I. poses to Google’s current business model. Meanwhile, Bing is enthusiastically leading the charge into the post-search era.
The emergence of Bing’s artificial intelligence marks the beginning of a new era for the Internet, where search may no longer be the primary means of finding information. The current design of the Web heavily relies on aggregation. I wonder what significance traditional websites will hold in a world where bots can perform the aggregation for us.
We are indeed living in the post-search internet, but let’s not forget that non-automated text or human-generated text will become a sought-after commodity.
This Chat GPT Lesson Plan and others are available FREE to all educators at edu.readyai.org
This article was written by Rooz Aliabadi (rooz@readyai.org). Rooz is the CEO (Chief Troublemaker) of ReadyAI.org
To learn more about ReadyAI, visit www.readyai.org or email us at info@readyai.org.
