ChatGPT vs Google: The New Battle For How We Get Our Information
Since November 2022, there has been noteworthy online chatter all over the internet regarding an AI chatbot, ChatGPT, developed by OpenAI. This conversational chatbot is a notable improvement from existing chatbots and could pose a challenge to Google, the search engine giant.
Google has enjoyed a stronghold with relatively no competition in the search engine domain for over two decades, creating the largest user base that anyone can trust and leveraging its swift innovation to stay ahead of its rivals. Nevertheless, a new rival has appeared on the scene in the form of OpenAI’s ChatGPT.
So, what exactly is ChatGPT, and how will it compete with Google’s search engine?
ChatGPT is a conversational AI chatbot developed by the American non-profit company OpenAI, an artificial intelligence research laboratory. The chat program is designed to converse with users naturally and in human language, using machine learning and language modelling algorithms (a Large Language Model) to understand and generate human responses.
ChatGPT is not a search engine like Google, but it has the potential to compete with Google to answer the searcher’s questions.
ChatGPT can answer various questions from coding to arts, provide explanations, and generate original content responding to user queries. In this sense, it is a more conversational and personalized alternative to a traditional search engine like Google, which gives relevant search results for websites.
However, it remains to be seen how successful ChatGPT will be in competing with Google, which has a vast infrastructure, huge data centres and years of user data. Ultimately, the success of ChatGPT will depend on factors such as its accuracy, efficiency, and ability to provide value to users that users can trust its answers.
It even impressed the Arvind Panagariya former vice-chairman of government of India think-tank NITI Aayog
Capabilities of the latest AI software is truly amazing. I asked ChatGPT to write a poem on Free Trade in the Alfred Tennyson style. See what it produced: pic.twitter.com/iiKcnTrHKQ
— Arvind Panagariya (@APanagariya) December 24, 2022
Tweet by Arvind Pangariya on ChatGPT
The big question is whether or not ChatGPT can actually compete with Google?
So far, the answer to the question isn’t so clear and simple. ChatGPT has shown reliable results in areas where training data for its GPT model is reliable.
ChatGPT is a chatbot that uses a Generative Pre-trained Transformer (so, the GPT in ChatGPT) model to generate question responses.
The ChatGPT is trained on an extensive data library, including Common Crawl, web texts, books, and Wikipedia. ChatGPT is designed to handle long conversations with users by keeping track of the conversation history.
Google’s search engine also uses a state-of-the-art sequential data method neural network model, RNN, to answer user queries. However, it is less sophisticated than ChatGPT’s GPT model. For example, Google’s search engine has to keep track of the user’s search habits to give relevant results, but search results are prone to produce wrong results.
Another difference is that ChatGPT can generate responses in multiple languages. In contrast, Google’s search engine can only generate results in English and limited languages like German, French, and Japanese. And can give results in language that can be limiting.
So, what will it mean for the future of search engines? ChatGPT and similar chatbots are promising contenders in the search domain. However, only time will tell if ChatGPT can live up to the hype.
What makes ChatGPT different from Google?
The below table shows the difference between ChatGPT and Google AI,
|ChatGPT is a chatbot designed for interacting conversationally.||Google is a collection of various algorithms, including text, image, video & ChatGPT, like conversational AI Google Assistant.|
|ChatGPT is based on the language model GPT, focused on Q&A-type conversational style exchange.||Google’s AI system can interact with users in a conversational style in Google Assistant and translate over 100 languages.|
|ChatGPT uses the GPT – Generative Pre-trained Transformer 3.5, a large language model.||Google’s voice search uses the RNN – Recurrent Neural Network.|
What can ChatGPT do?
▶️ Answer Questions
▶️ Create Content (based on prompts)
▶️ Content Rewriting
▶️ Write Code and Debug
▶️ It can take tests
▶️ AI Tutor in different styles
What can Google AI System do?
Answer Questions (Voice AI)
▶️ Image and voice recognition
▶️ Language Translations from over 100 languages
▶️ It can assist you in daily activities with a smart home
▶️ It can reverse image search
▶️ Google is a big recommendation system based on user’s behaviour
Limitations of ChatGPT
▶️ It can’t handle complex questions or queries and give wrong answers
▶️ It can’t understand natural Language
▶️ Only available in the English Language
▶️ It’s training data solely provided by Open.ai and can’t access the Web.
▶️ It can generate biased results towards women or minority groups
▶️ It’s not a commercial tool and still evolving, which makes it unsustainable if it isn’t commercialized.
▶️ It’s capacity to be used by many users is limited
Limitations of Google
▶️ Still, Google’s AI system gives wrong answers to complex or sensitive issues.
▶️ Its answers sometimes are biased and factless
▶️ It has limitations in recognizing the accent of various nationalities in the English language.
▶️ It is still an English-dominated search engine
▶️ Being a commercial search engine / AI system, it profiles the users, which compromises users’ privacy.
▶️ Disinformation and misinformation are still a major issues with Google search
▶️ It takes a lot of energy to generate the answers to users’ queries.
ChatGPT can eat Google’s Business Model!
Google’s business model is based on user searches on its search engine. It provides links to search results where you can see organic and paid links (above the organic links).
But in ChatGPT, you get immediate, accurate answers (sometimes it can be wrong) without leaving the ChatGPT’s user interface.
The following can be two ways ChatGPT can disrupt how we get our information.
Fact Based Answering Machine
As a language model with a database till 2021 and without internet access, it can be challenging to verify whether ChatGPT’s responses are fact-based, accurate, or incorrect. For example, it cannot distinguish between technical terms such as CPU and DNA.
For example, it’s not able to differentiate between CPU and DNA.
ChatGPT is sometimes amazing, and sometimes hilariously wrong. Here, it confidently explains to me why an abacus is faster than a GPU. 😃 pic.twitter.com/vpJZTStDnD
— Andrew Ng (@AndrewYNg) December 7, 2022
At the same time, ChatGPT can answer questions in any style, even emulating Einstein to teach physics. It makes ChatGPT an all-answering machine, which Google has tried to create over the past two decades with its knowledge graphs in web search and voice search in personal assistant space. ChatGPT’s ability to provide answers uniquely and engagingly could make the learning process more interesting and put it in direct competition with Google.
OpenAI’s new ChatGPT explains the worst-case time complexity of the bubble sort algorithm, with Python code examples, in the style of a fast-talkin’ wise guy from a 1940’s gangster movie: pic.twitter.com/MjkQ5OAIlZ
— Riley Goodside (@goodside) December 1, 2022
As an AI language model, ChatGPT is designed to learn and improve over time, becoming more refined in its understanding of natural language and concepts. Each new version of ChatGPT will result in more accurate and trustworthy answers. With every update, it will be able to provide answers based on a larger database, updated algorithms, and an improved ability to analyze and produce responses. Users can expect ChatGPT to provide increasingly reliable and insightful answers.
However, like any AI system, ChatGPT is not foolproof. In some cases, it may still make mistakes or provide insufficient or biased information. Nevertheless, ongoing updates and improvements will allow ChatGPT to evolve and provide accurate and trustworthy answers more effectively.
As shown above, ChatGPT can evolve into an AI tutor that can engage learners by imitating their favorite characters while speeding up their learning. But how does this affect Google? When we have questions, where do we go? We go to Google to search for answers. However, a conversational AI like ChatGPT can provide clear answers in the character of our choice. Learners may prefer ChatGPT over a search engine to learn and complete assignments.
However, this will remain a pipe dream if OpenAI can only capitalize on some of the inertia surrounding ChatGPT. As we have seen in the past, if technology is ahead of its time and requires more infrastructure to become commercially viable, it may not become a creative destruction that creates a new path for the future.