Is Google becoming vulnerable to AI Chat ChatGPT?


The new artificial intelligence tool, ChatGPT, powered by OpenAI, offers a new online search model. This chatbot responds to queries on computer programming and political science and gives detailed explanations. It has a question-and-answer format that allows users to drill down until satisfied.   

Until now, users conducting similar research on Google had to scan the search results and browse various websites to conclude. In contrast, ChatGPT delivers decisive answers within seconds. 

Threat to Monopoly 

Google so far has had a monopoly in search. Now a few companies founded by ex-Google executives think that it is time how search functions to change. The shift from the prevailing model of using keywords for search where search engines screened the web for specific terms to a new method of analyzing huge databases, understanding users’ queries, and providing rapid answers. ChatGPT uses these large language models to compose rapid-fire results. 

The large language tech breakthrough was initiated in some of Google’s research labs. However, entrepreneurs who left the company in recent years say that it will be difficult for Google to implement the new business model. The reason is Google’s current search results are accompanied by ads that are too lucrative to be disrupted. The last quarter saw Google Inc. earning $54.48 billion in advertisements, accounting for 78.9% of its total sales. The biggest driver was the search ads. 

Google will have to position itself at a higher bar when using new technologies than start-ups. ChatGPT and other similar models often provide incorrect information to queries in a convincing manner. Unlike a Google search, they also do not reveal the source of information, and OpenAI has admitted that they often provide incorrect answers. 

Google can direct users to sites that provide misinformation; therefore, it becomes risky for them to promote the same in their voice if they use a large language model. 

Authentic Information

As per Bloomberg news, OpenAI said that ChatGPT does not take information from the internet to provide answers to its users. It added that the current version is in research previewing mode and is trying to develop a safer and more reliable model intended to give answers accurately and not just advice. 

In recent years, Google has introduced combined searches using text and images for its users. A company such as Google is an ideal platform to conduct advanced AI work, given its access to massive data and computing infrastructure. Yet it has seen plenty of departures of its AI researchers as these engineers want to move faster, which is impossible at a big setup like Google. 

In 2017, Google contributed an influential paper, “Attention Is All You Need,” which introduced a system that helped AI models zero on the exact information from the data analyzed by using the concept of transformers.  

Seven out of eight authors of this paper left to start their own AI ventures, per their LinkedIn profile. Noam Shazeer, the founder of Character.AI, said that start-ups could move and start things faster. 

However, industry watchers say that the threat to Google is not worrisome. Morgan Stanley, however, acknowledged in a note to investors in December that users may be using such AI programs for travel and product reviews. 

Source: Bloomberg

About Author