ChatGPT, the natural language processing (NLP) model developed by OpenAI, has caused a sensation online since its Nov. 30 debut. The high level of buzz among tech enthusiasts, casual and professional alike, continued through the holidays as people experimented with the artificial intelligence-driven chatbot. The platform can be used to generate humanlike text to queries in real-time. And even our own writers and editors have explored the potential value of its use.

And now, Microsoft may plan to use ChatGPT to challenge Google in the search engine arena. 

Gradual Rollout for ChatGPT for Bing Search Engine?

The most prominent speculation around ChatGPT this week is that Microsoft is planning to add ChatGPT to its Bing search engine. A gradual rollout is expected as details about ChatGPT’s performance emerge as Microsoft is still weighing how the chatbot’s accuracy will impact its results. This will make Microsoft the first to respond to the potential impact on search engines. Microsoft already has formally partnered with OpenAI's image-generation AI platform, DALL∙E 2, announcing in October a planned integration with Microsoft's Azure OpenAI Service.

ChatGPT already has a leg up on Google search, so any partnership with Microsoft seems like a win over the Silicon Valley search engine giant. ChatGPT incorporates internet pages in its training set to summarize its query responses. Its summarized answers provide a more humanlike straightforward response that users have found convenient — vs. the numerous search-query results Google and others produce and internet users must wade through to discover content.

Further, search engine results appear alongside digital ads in Google's world whereas ChatGPT results display none. The ad-free chatbox in ChatGPT appeals to those who feel a bit overwhelmed with marketing messages.

Buzz among the tech community is that ChatGPT is a serious threat to Google’s search engine, which easily holds the No. 1 spot for search engine market share. Adding ChatGPT to Google’s chief competitor, Microsoft’s Bing, will certainly make that buzz louder and reinvigorate the fight for search engine market share between Bing and Google. Google has been the default search choice among consumers, dominating share to such an extent that it is the primary focus of SEO trends and tactics for thousands of businesses, marketing and content teams.

Related Article: OpenAI's New ChatGPT Might Be the First Good Chatbot

Is ChatGPT a Real Game Changer for Search Engines and SEO? 

ChatGPT has generated some pretty astounding results and responses since its launch and is causing many to rethink how they approach their online content optimization strategies. When I tried the platform, I asked for an explanation of Markov chain and received an example of code alongside a description. The same type of query in a search engine would have required me to look among several results on a search engine results page and then determine which results were useful to me.

If I'm querying a search engine with a phrase, I can look and see if the results fit what I'm trying to do. If I mention the phrase “apple pies,” I can look at the results on a search page to determine whether I want a recipe for apple pie or the information for a baker who can produce an apple pie. I can also look at the dates of a page and determine if the content is a fit.

ChatGPT sources from the internet, just like a search engine, but it is an artificial intelligence-based fusion of information. As a Reinforcement Learning from Human Feedback (RLHF) language model, ChatGPT, which uses the Generative Pre-trained Transformer 3 (GPT 3) language model, was developed using 570 gigabytes of text. According to the Stanford Institute for Human-Centered AI, this is 100 times more than the GPT-2 predecessor.

ChatGPT isn't perfect. It is formatted to provide the best summary in text boxes, rather than display sites ranked according to query relevancy and links to the sources. That summary can mask some details, making it difficult to know for sure if the information is neutral or overtly commercialized. The ChatGPT responses are not determining the value or even relevance of the "answers" it provides.

Despite the extensive training and much-lauded performance, OpenAI noted that ChatGPT operates on a knowledge base that ended in 2021. This means that some of the latest information is omitted in its answers. When I used ChatGPT to look at pipes in R programming, I noticed that it provided the older version of the pipe, which is from a library versus a new version which is incorporated in the latest versions of R programming. This is a minor example — many developers have not started to use the new pipe, and either version is an acceptable convention. But it does reveal how a user should know the nuances of queried topics, and whether any beta-like details matter in how helpful ChatGPT responses are. 

Related Article: ChatGPT's Impact on Customer Experience and Marketing

Learning Opportunities

What Can Marketers Gain From the ChatGPT Experience?

What's the key message here for SEO professionals and marketers? They should approach the hype around ChatGPT with measured evaluation rather than speculation. For instance, there are limits to how a ChatGPT user can rate the quality of the responses, whereas the results of a Google or Bing search query can be examined with much more nuance.

Not every platform is accepting the accuracy of ChatGPT. StackOverflow, a trusted Q&A site among developers, announced in December a temporary ban on any post content written by ChatGPT. The reason for the ban is to address concerns about how inaccurate details are being republished. On its site, StackOverflow states that “the average rate of getting correct answers from ChatGPT is too low.” Posts containing answers based on ChatGPT are “substantially harmful to users who are asking and looking for correct answers “the site noted. While people are still discovering the pros and cons of ChatGPT, it has become clear that ChatGPT could significantly influence content verification.

The takeaway? The world has officially entered an age of AI assistants where norms and standards need to be established sooner than later. So, marketers must watch closely and always continuously examine the source of data and information when considering campaigns and content.

Until now, AI assistants have offered benefits in confined environments — Apple’s Siri answers on your laptop or phone, and Amazon’s Alexa pulls its answers from the web and devices in its network for use at home. Other AI chat platforms designed for specific needs are becoming prominent, such as YouChat, Synthesia, Descript and Midjourney.

As these AI platforms achieve higher levels of adoption, people will alter their expectations over time. This in turn can impact downstream solutions such as analytics and refine what is measured to provide a better customer experience. SEO and search strategy will be included in any update.

Google's Not Sleeping at the AI Chatbot Wheel

Meanwhile, Google can’t ignore ChatGPT entirely. Blockbuster laughed at the idea of Netflix. However, Google is already showing signs that it’s ready to defend its market share. AI open-access journal published a piece that Google Research released a chatbot, co-developed with DeepMind, called MultiMed. The chatbot was developed on a public health language model trained on several extensive medical datasets to help professionals, researchers and consumers access information and answers to their medical questions.

This is a very niche application, limited to a specific industry and has nothing to do with the current Google search engine algorithm. But lessons learned can aid future development for Google’s other algorithm-based services, including search.

It's hard to say exactly how the elements that make up a search engine optimization will grow or diminish in value. The emphasis on content will certainly continue, while the treatment of links may be rethought.

A final lesson for marketers? Overall, marketers should not mark the arrival of ChatGPT as the demise of search engines and SEO. They should instead watch how customer expectations for chatbot-based services shift as awareness of ChatGPT grows into adoption by the general public.