OpenAI, the creator of Microsoft-backed ChatGPT, has confirmed that the chatbot can now search for current information on the internet.
Previously, the AI-powered system was only educated using data until September 2021.
As a result of the change, select premium users will be able to ask the chatbot questions about current events and gain access to news.
According to OpenAI, the capability will soon be available to all users.
Earlier this week, OpenAI stated that the chatbot will soon be able to converse with users using voice.
ChatGPT and other similar systems employ massive volumes of data to provide compelling, human-like answers to users’ questions.
They are likely to significantly alter the way people search for information online.
However, the “knowledge” of the viral chatbot has been locked in time until now. Its database was compiled from Internet content in September 2021. It could not access the Internet in real time.
So, for example, if you ask the free version when the last earthquake occurred in Turkey or if Donald Trump is still alive, it will answer: “I’m sorry, but I can’t provide real-time information.”
Some potential users have been put off by ChatGPT’s inability to take recent events into account.
“If this functionality or capability were not available, you would have to go to Google, Twitter or your preferred news outlet.” “You can now use this to get the latest news, gossip and current events,” says Tomas Chamorro-Premuzic, professor of business psychology at University College London.
“So the main implication is that it will absorb many of the incoming questions and queries that were going to the search engines or the media,” he explained.
However, Chamorro-Premuzic warned that using the platform for searches can be a double-edged sword.
“I think it’s good in terms of getting quick answers to your urgent and burning questions,” he added, but warned that the information offered through ChatGPT can be misleading without sourcing.
“If you don’t reliably indicate what the sources are, and you just do a mix and match of what’s out there… then the concerns are around accuracy and people just assume that the information they get out there is reliable when it is not. .”
US officials have already expressed concerns about the possibility of ChatGPT producing fraudulent data.
The Federal Trade Commission (FTC) wrote a letter to the Microsoft-backed company earlier this year, demanding details on how it addresses reputational dangers.
In response, OpenAI’s CEO stated that the company would cooperate with the FTC.
ChatGPT had not searched the Internet until recently for various reasons, including high computational costs. It is often said that each query costs OpenAI a few cents.
More importantly, the limited data served as a vital safety net.
ChatGPT was unable to begin regurgitating recently discovered dangerous or illegal material uploaded to the Internet in response to a query.
He could not spread falsehoods about politics or health decisions since he did not have access to them.
When asked why it had taken so long to allow consumers to search for updated information, the chatbot responded with three options.
He stated that building language models was time- and resource-intensive, that the use of real-time data could introduce errors, and that there were some ethical and privacy concerns about access to real-time information (in particular, copyrighted content). author) without authorization.
ChatGPT’s new features clearly reflect the important problem facing the AI field. To be truly useful, barriers must be removed or at least loosened, but doing so makes the technology potentially more harmful and vulnerable to misuse.
Read more: OpenAI’s ChatGPT can now talk, hear and see – a multimodal upgrade in history
Subscribe to our latest newsletter
To read our exclusive content, register now. $5/Monthly, $50/Yearly
Categories: Technology
Source: vtt.edu.vn