Why ChatGPT won’t replace search engines
Most likely, search engines should not be afraid of ChatGPT
SINCE OPENAI INTRODUCED the world of ChatGPT last November, people use it to code generation, writing poetry, creating discussions. Now, many are suggesting that people can adopt ChatGPT-style bots as an alternative to traditional web searches.
microsoft, $1 billion invested in OpenAI plans to release an implementation of its search engine bing, including ChatGPT until the end of March. According to a recent New York Times article , Google has red-coded it over concerns that ChatGPT could pose a serious threat to its $149 billion a year search business. Can ChatGPT really destroy a search engine mastodon like Google?
ChatGPT is good in that it – generates – but it is not a search engine.
It makes some sense that Google might consider chat, if it became widespread as a search tool, as a threat to its business model. If fewer people engage in organic search, this could potentially have a major impact on Google’s bottom line. But the real question here is: is it even possible to use ChatGPT for Google-like searches?
Let’s compare two technologies
Google and similar search engines are, strictly speaking, librarians. They can find any website you want and recommend other potentially related sites. It is then up to you to decide if these sites are what you are looking for.
On the other hand, ChatGPT is more like a psychic than a librarian. Its biggest drawback as an internet search tool is that it can’t connect to the internet – at least not to the internet after 2021.
This renders ChatGPT useless for most search queries. More often than not, Google searches are constantly related to breaking news and local or personal information such as places to eat, banking information, and websites—queries that require accurate and up-to-date information.
If you ask ChatGPT the same query in five different time sessions, you might get five different answers, possibly contradicting each other.
Add to this list of drawbacks the fact that chatbot language generation is incredibly slow compared to search engines. For example, when I entered into Google the query ” explain string theory ”, the query returned 134.000 results sorted by relevance in less than a second. In ChatGPT, this takes at least a minute.
Even if chatbots could respond faster, their paragraph format responses are not exactly optimal for readers who want to process a large amount of information quickly. Once the novelty of interacting with a robot wears off, most people will likely go back to being informed as usual.
Most likely, ChatGPT style bots will be merged with existing search engines to offer a user interface that will serve both traditional search engine queries and chatbot prompts. This is the model adopted You.com , a specialized search engine that launched its own GPT-like chatbot in December. Instead of replacing the traditional You.com search, the new YouChat feature simply appears as a link below the search bar. The innovation here is to put two very different AI apps on the same page. It’s probably safe to assume that Microsoft will do something similar when it integrates ChatGPT into Bing this spring.
ChatGPT is a very interesting bot and with further development it may even have a bright future as a digital assistant. But this does not mean that chatbots will revolutionize search engines, much less replace it.
At the moment they are too slow and sluggish. Sure, there is some overlap in users, but the technologies serve completely different purposes.
