Reading Time: 16 minutes -

[Editor’s Note: 19.05.23] A lot has changed since our original article dated 18th January 2023, which discussed ChatGPT vs Bing and what an AI search engine might look like. In only four months, we’ve seen an almost unprecedented amount of change, with new providers coming onto the scene, including Microsoft’s Bard and OpenAI’s GPT-4 with web browsing and plugin availability as part of ChatGPT Plus.

The question is now more of ChatGPT vs Bard vs Bing and who will win the race to become the first true AI search engine. But more than that, we have other questions to consider, such as: Will open source models play a role (e.g. Meta’s LLaMA)? Will current providers like Google, OpenAI and Microsoft turn to open source? Will a new contender emerge as the winner, perhaps as Anthropic’s Claude? Will search engines as we know them even exist, or will internet-enabled AI chatbots replace them altogether? The list goes on… This update gives our original article a major refresh and a new take on the topic.

Introduction

The world of search engines is undergoing a significant transformation, with artificial intelligence (AI) playing a crucial role in driving this change.

AI platforms like OpenAI’s ChatGPT, Microsoft’s Bing, and Google’s Bard are revolutionising the way we search for information and interact with technology. As technology giants race to leverage AI in their search engines, users can expect a new era of personalised and efficient search experiences. In this article, we will explore the concept of AI-based search engines, the key players in the market, the primary differences between them, and the potential implications of these AI-powered search tools.

Should Google be concerned or excited?

If you would prefer to listen, check out our original Soundcloud podcast.

Let’s dive in and find out!

AI search engine: the emergence of a new era

What is an AI search engine?

AI search engines use machine learning algorithms to understand and process user queries, providing more relevant, accurate, and personalised results. These algorithms are designed to learn from vast amounts of data and improve their performance over time, leading to better search outcomes for users.

They rely on Large Language Models (LLMs) to predict and generate human-like responses. These LLMs analyse the statistical properties of language and make educated guesses based on previously typed words. The ability to generate plausible-sounding statements makes AI search engines extremely powerful, but it also poses challenges, as these generated statements may not always be accurate or reliable.

The rise of AI-powered search engines

The popularity of AI-powered platforms like ChatGPT has disrupted the traditional search engine landscape. With the integration of AI technology, search engines can now provide more personalised, accurate, and relevant search results, transforming the way users find information online.

ChatGPT, Bing, and Bard are some of the leading AI platforms that are pushing the boundaries of search technology, providing users with more intelligent, conversational search experiences. We’ll explore these in more detail below.

The role of Large Language Models (LLMs)

AI search engines are primarily powered by large language models (LLMs) like OpenAI’s GPT-3, Google’s PaLM 2, or Microsoft’s Copilot. These models can understand and generate human-like responses, opening up a range of possibilities in various fields, from customer service to language translation.

These LLMs are trained on vast amounts of data and can predict the next word in a given sentence, making them capable of generating coherent, contextually relevant text.

What are the benefits of AI-based search engines?

AI-powered search engines can offer several advantages over traditional search engines:

Personalised and contextual search results

AI-based search engines like Bing Chat and Google Bard have the potential to reshape the online search landscape by providing more personalised, contextual search results. By leveraging AI technology, these search engines can better understand user intent and deliver more relevant, accurate search results, ultimately improving the overall user experience.

Conversational and interactive search

One key advantage of AI search engines is their ability to offer conversational, interactive search experiences.

Users can engage in a natural, back-and-forth dialogue with the search engine, allowing for more nuanced and precise search queries. This conversational approach has the potential to streamline the search process, enabling users to find the information they need more quickly and efficiently.

Integration with other AI technologies

As AI search engines continue to evolve, they are likely to become increasingly integrated with other AI technologies, such as virtual assistants like Siri and Alexa. This integration could lead to even more seamless, intelligent search experiences, providing users with a unified, AI-powered ecosystem for finding information and interacting with technology.

What is ChatGPT? What’s the big deal?

Since ChatGPT was recently made public for testing, it has taken the internet by storm. From answering questions to helping beginners fix CSS and coding problems, ChatGPT frequently provides help and its conversational style is almost indistinguishable from a human.

This isn’t your ordinary chatbot-tier AI we’ve seen a thousand times before. It’s something quite unique and very exciting.  As regular readers will know, we’ve covered ChatGPT and GPT-3 in great detail recently, focusing mainly on AI SEO, content ideas, and even how to detect AI, but this time we’re considering how the news of a ChatGPT Bing partnership could change the search landscape forever and potentially be the driving force behind a new breed of AI search engine.

It talks like a human would

ChatGPT is an intelligent AI conversational chatbot that handles queries in a similar way to how a human would. Users can ask follow-up questions relating to the initial inquiry, and it will remember the conversation thread. These types of contextual-based answers are what make ChatGPT particularly exciting.

For example, you could ask the AI to write you a paragraph talking about the benefits of daily exercise. If the reply sounds good but is a little too lengthy, you can simply reply with “make more concise” and ChatGPT will do just that as it remembers the previous parts of the conversation.

What makes ChatGPT particularly impressive is its ability to not only write informative replies but also do it in the style of a particular person. We decided to put this to the test by having the tool create a recommendation in the style of Donald Trump due to his distinct and brash personality.

From text to code, it can solve specific problems

You can also be very specific when looking for suggestions and recommendations. Many users have even been using ChatGPT to solve coding problems. This type of usage is particularly ground-breaking as these types of problems are so intricate, Google can’t really help much by way of a conversational search at the moment or help with these types of problems. ChatGPT, however, can.

Obviously, this AI is very impressive as a chatbot, but it also has huge implications for search as well as virtual assistants. Read on to find out more.

What are the limitations of ChatGPT?

As impressive as ChatGPT is, there are still some limitations that users frequently encounter.

Restricted use and limited knowledge of events 

The AI is trained to decline inappropriate requests and has limited knowledge of the world and events after 2021. It is worth keeping in mind that ChatGPT is still in its testing phase meaning OpenAI is subject to remove or add new features when the final release rolls out.

Character limits and bandwidth

Another limitation is that there is a character limit imposed for queries and replies. Some users have gotten around this problem by having the AI write longer posts in multiple parts due to the fact it can remember previous parts of the conversation.

Bandwidth is also a concern. As ChatGPT is a hot topic at the moment, the chat servers are being hit hard resulting in regular downtime and error messages.

limitations can be imposed

Keen-eyed users of the ChatGPT subreddit have been noticing a degraded level of intelligence since a recent update was made live. This is likely the result of restrictions and limitations being imposed due to some users using it to create controversial content and hate speech.

With the prevalence of cancel culture and ChatGPT being a hot topic at the moment, it makes sense that the OpenAI team would be cautious and reluctant to allow their AI to be completely unrestricted, especially after Microsoft’s Tay chatbot posting racist rants and images via Twitter some years ago.

There is certainly a delicate balance to be struck when it comes to limiting ChatGPT. Hopefully, OpenAI will continue to fine-tune things in order to keep everyone happy. It is worth keeping in mind that as of the time of writing, it’s still in a trial phase where the objective and goal is to learn how users interact with the chatbot.

Microsoft takes a giant leap to tackle Google’s dominance 

The fundamental way search results have been served to users through web browsers has been largely unchanged for many years. We’ve had updates within Google such as featured snippets and many algorithm changes to enhance relevancy, but very little has come via fundamental overhauls or improvements.

Google currently has over 80% of the market share when it comes to search, with Kinsta reporting 86-96%, but Bing is slowly increasing their presence.

ChatGPT Bing (now known as Bing Chat)

Sitting just below 10% isn’t where Microsoft wants to be. It was recently announced that Microsoft would be integrating ChatGPT into Bing’s functionality in order to enhance its search capability from the end of March 2023.

Conversational, follow-up style search

Back in January 2023, we predicted that ChatGPT Bing AI search engine would make search more accurate by allowing follow-up questions to the initial enquiry:

“If ChatGPT is anything to go by, it could also provide answers faster as traditional search often requires checking multiple sources and scrolling through web pages to find an answer that’s relevant. It also has the ability to provide opinions, summaries, and answer queries about programming code in a way that’s simply not possible with traditional search engines right now.”

We continued with:

“This will likely result in Google integrating similar changes in the near future, but will they be able to keep up with ChatGPT’s momentum? A first implementation of a tool such as ChatGPT could look something like this within Google’s search. Google may introduce a secondary search bar that allows users to continue the conversation thread after their initial enquiry. This would allow an initial question to be asked such as “How far away is the moon from the earth” and a follow up question such as “When was the last time we went there”. This means that the user doesn’t have to type out a full explanation of the query, but a continued search.”

Bing vs Google: the battle of the AI search giants 

The race to dominate the AI search engine market is primarily led by two technology giants: Google and Microsoft, but now also OpenAI. Whilst this article is mainly focused on the future of search and the experience provided to users, we’ve also explored Bing SEO vs Google SEO in our article below. For more information on the past, present and future of SEO and search engine ranking factors, it’s worth a read.

Bing SEO vs Google SEO

Bing Chat: Microsoft’s AI-powered search engine

Bing Chat has gained considerable attention since incorporating new AI technology from OpenAI, making it a formidable competitor to Google’s search business. Microsoft’s Bing, with the integration of ChatGPT, offers users a more interactive and conversational search experience, enabling them to ask questions and receive AI-generated responses. The incorporation of AI had increased Bing’s popularity, although it has returned to previous levels after an initial surge.

Google Bard and Magi: Google’s answer to AI search 

Since ChatGPT was released, Google has issued what some have referred to as a ‘code red’ with emergency meetings being called to assess the threat of ChatGPT to its business model and search as a whole.

Google has been working on its LLM called LaMDA, which stands for ‘Language Model for Dialogue Applications’, for quite some time. According to Google, LaMDA can “engage in a free-flowing way about a seemingly endless number of topics”. More recently, Google has announced a new LLM called Pre-training with Abstracted Language Modeling-2 (or PaLM-2 for short) that will become the underlying technology for its new AI tools.

It is worth remembering that Google is no stranger to AI and conversational chatbots. ChatGPT (the GPT of which stands for Generative Pre-trained Transformer) was built on top of GPT-3’s family of large language models that are built on Transformer — an open source neural network architecture that Google made available in 2017.

In response to the latest threat from AI competitors like Open AI and Microsoft, Google has developed its own AI chatbot platform called Bard, which is currently based on LaMDA. As Bard moves to PaLM-2, it’s expected that this will bring third-party extensions to Bard.

Part of the Magi project, Bard aims to provide a more personalised, user-friendly search experience, leveraging the company’s vast knowledge base and AI capabilities. However, Bard is just one element of the overall Magi project. While the exact launch date is unknown, Magi is expected to enhance Google’s current search engine with new AI features, including more accurate search results and an improved understanding of user intent.

ChatGPT vs Bard: a brief comparison

Technology and applications

ChatGPT and Google Bard both rely on LLMs to generate human-like text responses. However, there are differences in their underlying technologies, with ChatGPT powered by OpenAI’s GPT-3 and GPT-4, but Google Bard powered by Google’s LaMDA (soon to become PaLM-2). These differences may lead to variations in the quality of search results and user experience provided by each AI search engine.

AI-powered features and capabilities

ChatGPT and Google Bard both offer a range of AI-powered features and capabilities to enhance the search experience. These include generating summaries, answering complex questions, and providing contextual, conversational follow-ups. However, their respective implementations and effectiveness may vary, leading to differing user experiences.

Open source vs proprietary models

As covered in our article here, there’s a possibility that open source may actually lead the way. An internal communication from a Google staff member, which was unintentionally made public, has admitted that the tech giant is facing challenges in rivalling open source artificial intelligence systems.

The document underscored the swift advancement of open source projects and their benefits, including affordability, adaptability, and easy access.

Could Meta (Facebook) be a contender?

Meta’s LLaMA, a notable open source project, has drawn a sizable number of developers and has been a focal point in AI research discussions. It has laid the groundwork for countless open source AI initiatives, showing the power of open source to spark creativity, match up to, and even outdo proprietary AI systems.

Open source AI: Google vs OpenAI vs Meta 

Whilst not strictly open source, the open source beggings and nature of OpenAI provides several advantages for its development and adoption.

By allowing developers and researchers to access and modify the underlying AI model, open source projects like LLaMA can benefit from a broader range of ideas and innovations, ultimately leading to more robust, user-friendly solutions. Furthermore, open source projects can facilitate greater transparency and collaboration, helping to address potential concerns related to AI bias, misinformation, and other ethical issues.

Right now, it’s uncertain what role Meta or LLaMA will play in search (if any), but it’s clear open source has Google worried and there are even talks of OpenAI releasing an open source model in the near future.

Other possibilities?

Anthropic’s Claude: A new AI search engine contender?

Anthropic’s Claude is another AI platform that has recently emerged as a competitor in the AI search engine space. While not as well-known as ChatGPT, Bing Chat or Bard, Claude offers a promising alternative for users seeking a more personalised, context-aware search experience and may end up being a competitor.

What about OpenAI’s ChatGPT, could this become a search engine? 

The version of ChatGPT developed by OpenAI is yet another possibility. With a partnership with Microsoft, ChatGPT has become a significant threat to Google’s search engine dominance, prompting Google to invest in AI product development for search.

However, could ChatGPT itself before an AI search engine contender? Well, it already is. While Google make plans to incorporate more AI capabilities into the Google search engine through the Magi project, ChatGPT Plus users currently get access to GPT-4 with web browsing built in. What is this if not an AI search engine?

Much like Bing Chat and Bard, all three platforms now provide access to the internet and the ability to search the web with the assistance of their AIs, so as well as being chatbots, they are all arguably search engines as well.

AI integration into existing search engines or a new breed altogether?

Are search engines already powered by AI?

It’s important to realise that AI has been playing a role in how search engines rank content for a long time.

For instance, Google uses AI algorithms like RankBrain to understand and process user queries better, providing more relevant search results. Although AI technically plays a role in the search algorithm rather than the search engine (the user interface to search), it’s certainly not a new concept in the world of search, with AI helping to:

• Improve query understanding and processing
• Enhance search result relevance and accuracy
• Provide personalised search results based on user behaviour and preferences
• Offer better language understanding and processing

Will AI be integrated into existing search engines?

We are already seeing Bing Chat being integrated into the standard Bing search engine experience.

Google is investing heavily in AI research and development to protect its search engine dominance. The company is reported as focusing on two main projects: the development of the Magi search engine and the integration of AI features into its existing search engine.

Google plans to introduce Magi’s features to an initial user base of one million, which will gradually increase to 30 million by the end of the year exclusively in the United States.

This means that the Google search engine that we all know and love will be enhanced with AI features from the Magi project. This includes more accurate search results, an improved understanding of user intent, and the ability to personalise search results based on user behaviour and preferences.

Will AI chatbots replace search engines?

Given the unprecedented surge in popularity of ChatGPT and the changes already available with ChatGPT Plus providing GPT-4 and web browsing capabilities, it’s possible that browser-enabled chatbots like ChatGPT, Bing Chat and Bard may replace search.

Taking ChatGPT as the best example right now, it has some of the most advanced AI capabilities available via GPT-4, access to real-time web browsing, a clean and easy-to-use interface, a dedicated mobile app, a plugins directory, and significant integration with other platforms.

With so much going for it, you might wonder why anybody would choose to go to a traditional search engine.

The most likely outcome

The most likely outcome is that we will see a mixture of the above options come true.

Regardless of whether Google, Microsoft or another party wins the race, it’s very possible that the terms ‘AI search engine’ and ‘AI chatbot’ will become synonymous with each other.  Ultimately, you can expect to see all major search engines providing more interactive AI capabilities in a chatbot-like fashion.

Blurring the lines between search engine and AI assistant

The Bing and ChatGPT partnership provides Microsoft with the opportunity to offer deeper integration of search within devices such as the Surface Duo.

ChatGPT could take on AI assistants like Alexa and Siri

If Microsoft integrates AI at an OS level, ChatGPT could become a far superior voice assistant than anything else available and put the likes of Alexa and Siri to shame. This would be a huge selling point for their devices which have frequently struggled to gain traction with audiences.

Integration between AI, search engine and web browser

Alternatively, providing ChatGPT within a Bing and Microsoft Edge-powered smartphone app would be a huge incentive for users to switch from their regular web browser.

Currently, there is no real incentive to change browsers, all of your bookmarks and shortcuts are there is little to motivate a change. If Bing offered something like ChatGPT functionality within its browser natively, this would surely justify Firefox and Chrome users making the switch.

Not a fan of Bing? This is still good news — “A rising tide lifts all boats”

A lot of users are loyal to Google, and won’t be moving to Bing even with ChatGPT integration. You may wonder what the benefits will be for you as someone that intends to remain using Google search for the foreseeable future.

A big positive of Bing striking a deal with OpenAI is that it will encourage search providers and AI assistants alike to improve upon their existing technologies.

Improvements and advancements of assistants such as Siri have been slow in recent years. Hopefully, the success of ChatGPT will encourage the likes of Apple and Google to take another look at how useful their current assistants could be with an upgrade to how they function.

ChatGPT integration into phones via an assistant technology could transform the way we use our smartphones. A change like this would mean far more intelligent assistants than the ones we have at present. For example, we could ask the AI to schedule meetings, text everyone invited with the details, and send location data to the attendees all at once in just a few seconds. There is so much potential here when it comes to the future of AI search and voice assistants within the mobile phone and smart device market.

What impact could have on the digital economy?

As AI search engines become more prevalent, they have the potential to significantly impact the digital economy.

Websites and content creators that rely on traditional search engines may need to adapt their strategies to stay relevant in a world dominated by AI-based search engines. Websites that rely on subscriptions and ads for revenue may be affected if users no longer need to visit their sites to find answers to their questions.

This could radically change how businesses and marketers approach SEO and PPC for example, forcing them to find new ways to get found using AI.

ChatGPT vs Bing vs Bard: what’s the verdict?

The rapid advancements in AI have significantly the technology industry, with the search engine market being no exception.

At the current time, it’s hard to position open source, platforms like Meta’s LLaMA or new contenders like Claude in the race. It’s too early to anticipate what role they will play.

It’s challenging to definitively declare a “winner” in the ChatGPT vs Bing vs Bard (or OpenAI vs Microsoft vs Google) debate, but we do have some hunches.

Each platform offers its own unique set of features, capabilities, and underlying technology, catering to different user preferences and requirements, meaning they’re all likely to be around for some time to come. However, based on our own experiences using ChatGPT, Bing and Bard, we would give OpenAI and ChatGPT our thumbs up.

We favour OpenAI and ChatGPT but this is for a number of reasons, not limited to:

• They were ‘first to market’ with a freely available generative AI chatbot
• They have grown faster than any app in history and already have a huge customer base
• They have the power of publicity and popularity on their side, with huge followings and communities on social media
• The answers are generally higher quality than those from Bard and Bing Chat
• The level of programming code expertise and support is exceptional
• Integration with other platforms and third-party plugin support is already available
• Developers have had API access to OpenAI’s models for some time, which now includes GPT-4

An important caveat to make here is that the above opinion is based on using ChatGPT Plus with GPT-4 and web browsing. As a free ChatGPT user, we would probably favour Bard with its access to the internet, and as lovers of open source, we would like to see this win the AI search engine race overall.

Conclusion: the race to AI search engine dominance

While AI chatbots and AI-powered search engines continue to evolve, conventional search isn’t going away any time soon.

Given the current trajectory of AI development and the investments made by major technology companies, it’s difficult to accurately predict which player will ultimately dominate the AI search engine market. Google, Microsoft, and OpenAI all have significant resources and expertise at their disposal, making it a closely contested race.

The rate of change in the world of AI is unlikely anything seen before. As AI search engines continue to evolve and improve, users will likely benefit from an increasingly intelligent, personalised, and efficient search experience. The reality is that it’s still very early days and anything is possible, including the unknown. That said, the competition between ChatGPT, Bing, Bard, and other AI search engines will only serve to drive innovation and enhance the overall user experience in the years to come.

Whoever wins, it’s clear that AI-based search engines will play a crucial role in shaping the future of online search, providing users with more personalised and efficient search experiences.

What do you think about this news? Do you think Bing’s integration with ChatGPT could spell the end of Google search and the beginning of a new breed of AI search engine? Let us know in the comment