Open AI, the Microsoft-backed creator of Chat GPT, has confirmed the chatbot can now browse the internet to provide users with current information.
The system relied on the training data available up to September 2021 which is also powered by artificial intelligence.
As a result, some upper-class customers can query the bot about contemporary events as well as retrieve news.
The company, Open AI, said that, the feature will be opened for all users very shortly.
Moreover, Open AI has earlier in the week announced that they would soon upgrade the chatbot to allow voice communication between the users and it.
Chat GPT along with other systems, make use of large quantities of data which aid in generating believable responses for a user’s question.
The virtual chatbot is supposed to revolutionise how people find information on the internet. However, the bot’s knowledge remains dated until now. In fact, its database came from a compilation of content that was present on the internet in September 2021. It could not do real time browsing of the net.
For instance, enquire from a free version when did the latest earthquake strike on Turkey, or is America’s president still alive and it replies “my apologies; I do not offer up-to-date information”.
A number of people have not warmed up to it due to its failure to incorporate new developments.
According to Tomas Chamorro- Premuzic who is a professor of business psychology at University College London, “without these functionalities and capabilities, you would have to rely on Twitter, the news outlets, Google.” He continued by saying that now, the platform could be considered to
“The main implications are going to be soaking up lots of questions and queries which were going to go into the search engines or the news houses he added.
However, Mr. Chamorro-Premuzic emphasized the dual nature of using the platform for searches.
“It’s beneficial for quickly addressing urgent questions,” he acknowledged. Nevertheless, he cautioned against the lack of proper sourcing, noting that information from ChatGPT could be misleading without clear references.
“If it doesn’t transparently indicate reliable sources and instead mixes and matches existing content indiscriminately, there are concerns about accuracy. People might assume the information is trustworthy when it’s not,” he warned.
Open AI has already faced scrutiny from US regulators regarding the risk of Chat GPT generating false information. Earlier this year, the Federal Trade Commission (FTC) sent a letter to the Microsoft-backed company, seeking information on how it addresses risks to individuals’ reputations. Open AI’s CEO assured that the company would collaborate with the FTC in addressing these concerns.
Several factors contributed to the delayed integration of internet searches into Chat GPT. One factor was the considerable computing cost, with each query reportedly costing Open AI a few cents. However, the more significant reason was the intentional limitation of data access, providing a safeguard against the regurgitation of harmful or illegal content newly uploaded to the internet in response to user queries. This limitation prevented Chat GPT from spreading misinformation planted by malicious actors, especially regarding politics or healthcare decisions.
When asked about the prolonged time it took to enable users to access up-to-date information, Chat GPT provided three explanations. It highlighted the time and resource-intensive nature of developing language models, the potential introduction of inaccuracies with real-time data, and expressed privacy and ethical concerns, particularly regarding accessing copyrighted content without permission.
The introduction of Chat GPT’s new functionalities underscores the challenging dilemma faced by the AI sector. Striking a balance between utility and safety poses a significant challenge, as removing or loosening guardrails may enhance the technology’s usefulness but also increases the potential for misuse and dangers.
Leave a Reply