OpenAI, the creator of ChatGPT sponsored by Microsoft, has confirmed that the chatbot can now browse the internet to provide users with current information. Previously, the artificial intelligence-powered system was trained using only data through September 2021.
Some premium users will now be able to access news and ask the chatbot queries about current events.
OpenAI stated that the feature would shortly be accessible to all users. OpenAI disclosed earlier in the week that the chatbot will soon be capable of voice conversations with users.
ChatGPT and other similar systems use massive quantities of data to generate responses to user queries that are convincingly human-like.
It is anticipated that they will significantly alter how people seek information online. But until now, the knowledge of the viral chatbot has been suspended in time.
Its database is composed of the internet's content as it existed in September 2021. It was unable to access the Internet in real time.
If you query the free version, for instance, when the last earthquake struck Turkey or if Donald Trump is still alive, it will respond that it cannot provide real-time information. The inability of ChatGPT to account for recent events has deterred some potential users.
According to Tomas Chamorro-Premuzic, professor of business psychology at University College London, without this feature or capability, one would have to visit other websites or their preferred news source for the latest news, rumours, and current events. However, you can now use this as a source for such information.
Mr. Chamorro-Premuzic added, however, that searching on the platform could be a double-edged sword. "I think that is a good thing in terms of getting quick answers to your pressing, burning questions," he said, but cautioned that information provided via ChatGPT could be misleading if not sourced.
OpenAI has already come under scrutiny from US regulators due to the possibility of ChatGPT producing fraudulent information.
The Federal Trade Commission (FTC) sent a letter to the Microsoft-backed company earlier this year, requesting information on how it addresses reputational risks.
In response, the CEO of OpenAI stated that the company would collaborate with the FTC.
The cost of computation was one of the reasons why ChatGPT did not search the internet until now. It is commonly stated that each query costs OpenAI a few cents. Nonetheless, the limited data provided an invaluable safety net.
In response to an inquiry, ChatGPT could not begin regurgitating harmful or illegal content it happened to discover newly uploaded to the Internet.
It could not spread false information about politics or healthcare decisions inserted by bad actors because it lacked access to it.
When asked why it took so long to enable users to search for current information, the chatbot provided three responses.
It was stated that developing language models was time-consuming and resource-intensive, that using real-time data had the potential to introduce inaccuracies, and that accessing real-time information - particularly copyrighted content without permission - raised some privacy and ethical concerns.
The new capabilities of ChatGPT perfectly illustrate the enormous dilemma confronting the AI industry.
To be truly useful, the guardrails must be removed or at least loosened, but this makes the technology potentially more hazardous and susceptible to abuse.