ChatGPT is an advanced language model capable of understanding and generating natural language. It has gained instant success.
Simultaneously, The New York Times says that Google is concerned about the recent debut of ChatGPT. Earlier in December, Google officials worried that rushing to release AI technology could damage the company's reputation, despite its massive investment.
The main flaw in the claim that Chat GPT can replace Google is that the latter is a simple search engine while the former is a complex machine-learning system. From a technical standpoint, comparing ChatGPT and Google is like comparing oranges and apples. However, we can use them for search purposes that are fundamentally different in character.
How ChatGPT Works
ChatGPT takes advantage of massive language models. Imagine a more sophisticated form of the already available chatbots. Because of its knowledge of the natural language, ChatGPT can comprehend the meaning behind the questions being asked. To give just one example, it can tell the difference between a question and a statement.
The system has been trained on 45 terabytes of data, the equivalent of 175 billion sentences. As a result, its conceptual grasp is quite broad. For example, if you ask for a pie recipe, it will know what you're talking about because it has previously scanned and analysed the recipe. Because of this, it is often misunderstood as a search engine replacement.
Furthermore, ChatGPT is like a genius kid who reads every book and every public social media post and answers questions: For example, requesting dating advice from an AI when machines do not experience human love. You will receive a textbook answer, but never one based on experience.
Where Search Engines Score
The main issue is that ChatGPT's information needs to be more accurate. It contains data that has been trained to a specific date. Consequently, ChatGPT has yet to learn of the world after that date. First of all, it doesn't crawl the web. Therefore it can't find the required data and reveal its sources. Second, the information used to train it, known as the "knowledge base," will be obsolete, as it is trained only till 2021.
A search engine, in comparison, is more akin to a librarian who has access to every book ever written. And it can bring back the relevant library with just a question.
The information displayed in ChatGPT is organised into paragraphs. It is a summary of the information available on the topic. In contrast, the information returned by a search engine is presented as tables or lists, which can run into several pages. It contains connections to publications and resources that may provide an answer to your inquiry.
Due to lack of proper sourcing, ChatGPT may occasionally provide inaccurate and unintelligible data. When baffled by the questions, it resorts to guessing rather than asking for clarification. However, it can reject unsuitable questions, but this capability is still developing.
Competition Plays Catch Up
Google foresaw the world's direction and debuted its chatbot in 2021. The abbreviation for Google's technology is LaMDA, which stands for Language Model for Dialogue Applications.
"LaMDA's verbal abilities have taken years to develop. Like several recent language models, such as BERT and GPT-3, it's built on Transformer, a neural network architecture that Google Research designed and released in 2017," Google noted at the chatbot's introduction.
Similarly, Microsoft is developing a version of their search engine Bing that utilises the artificial intelligence powering ChatGPT, which OpenAI developed. This feature could make Bing more competitive against Google's search engine at a time when Alphabet's advertising revenue growth has slowed considerably.
Although Google's AI research is believed to be on par with other notable tech companies, it has only tested software with extremely tight boundaries. For instance, the AI Test Kitchen app provides access to image and text production tools comparable to OpenAI's DALL-E and ChatGPT. However, Google restricts users' queries to these programmes significantly.
A Limited Search Engine
ChatGPT simplifies and expedites the consumption of information. It could be more effective as a search engine for questions requiring greater precision and depth. It cannot get you a doctor, cannot provide dating advice, cannot display current news happenings, and is not 100 per cent accurate. In the era of fake news, accuracy is essential, and distinguishing between facts and views is critical for human decision-making.
So, can ChatGPT, therefore, replace search engines? The answer is NO. ChatGPT cannot currently be used as a substitute for search engines. It is because chatbots like ChatGPT still need to gain the trustworthiness necessary to replace search engines altogether. Although ChatGPT can be incredibly inventive, it can only sometimes be factually accurate. It might produce data that needs to be updated and relevant.
Nivash Jeevanandam writes stories about the AI landscape in India and around the world, with a focus on the long-term impact on individuals and society. Views are personal and do not represent the stand of this publication.
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
Find the best of Al News in one place, specially curated for you every weekend.
Stay on top of the latest tech trends and biggest startup news.