ChatGPT – Super Artificial Intelligence can also provide inaccurate information

Tram Ho

This is an artificial intelligence chatbot developed by the OpenAI Startup Company (USA) and according to information from CNBC (Consumer Business and News Channel), Microsoft will invest 10 billion USD to continue. develop this product.

Already, millions of users are testing this artificial intelligence (AI) to write articles, write poetry, and even advise on love stories. However, recently experts also warned, ChatGPT can give incorrect answers, so users must be careful with these artificial intelligence machines.

In a conversation with Google (Talk at Google), Sam Altman – Founder of Open AI Company said that he studied computer science at Stanford University and always wanted to create safe artificial intelligence. being friendly can help people. After many years, Sam is still pursuing research in artificial intelligence and 2 months ago, Sam’s Open AI Company introduced ChatGPT, with a beta version.

ChatGPT - Siêu Trí tuệ nhân tạo cũng có thể cung cấp thông tin không chính xác - Ảnh 1.

ChatGPT is understood as a computer system using artificial intelligence technology, which can answer questions in many different human languages. (Artwork: KT).

Not only attracting the attention of the technology world, after 5 days of launch, ChatGPT has reached a record of 1 million users and quickly reached 100 million users. Sensor Tower’s statistics show that, to reach 100 million users, the Google Translate application needs 6.5 years, Instagram takes about 2.5 years, Tik Tok application needs 9 months while the record level of 2 months is already in the about this Chatbot.

ChatGPT is understood as a computer system using artificial intelligence technology, which can answer questions in many different human languages. Even, from simple keywords, ChatGPT can generate text just like humans write, so in the past 2 months, the keyword “ChatGPT” was the most searched on Google.com.

Mr. Nguyen The Trung – Chairman of the Board of Directors, General Director of DTT Technology Joint Stock Company – said: “We can say it is an evolution from Google’s search engine. ChatGPT can exchange in the form of language is very much like a person talking to us, with that it is possible to understand a person’s question given context as well as meanings.

Then organize the information that is available to create answers that are coherent, understandable and continuous in the flow of communication between the two parties. I chose to ask ChatGPT, “Can ChatGPT counter misinformation?”, to which ChatGPT replied, “ChatGPT is a language model trained on large amounts of data, but it is not capable of ability to accurately respond to false information. ChatGPT will respond according to the data and information it has learned, but there may be cases where the answer is wrong or incorrect. It is therefore always best to check and confirm information with reliable sources before using or applying it.”

Machine Learning – Machine learning is making Artificial Intelligence (AI) smarter, so when the data is searched, filtered and sent to ChatGPT by the machine, the accuracy will depend on the source. information available in the computer system. That is also the reason that ChatGPT is not completely sure whether its reply information is correct or not? Therefore, even with Sam Altman – the founder of this Chatbot, he also wants users to be careful with artificial intelligence technology.

According to Machine Learning: “We released the technology much later than people would have liked. I also doubted whether ChatGPT needed more than one upgrade? When introducing the paid version, we also hope more people use it and gradually GPT Chat will get better Because artificial intelligence can learn on its own, so then it gets better, every year it gets a little better, next year will be better than last year.

That also needs users to always be careful with the machine, especially when it is getting smarter. I think, at first use you find it great, but after 10 times, then 100 times of use, you find that sometimes it will give wrong results. However, as we all know if the data is not correct, the results will be wrong – just like the fake information that still appears in cyberspace.”

It is estimated that an average of 13 million people use ChatGPT every day and on February 2, ChatGPT announced a fee version, with a fee of 20 USD/month. Currently, ChatGPT can be used with IP addresses and phone numbers in the US, Canada, and some other European countries.

In Vietnam, there are many forums sharing how to register an account to use ChatGPT, but experts also warn about the risks that users may lose money, but still cannot use it. or be led to fraudulent links, stealing personal information.

Therefore, with any intelligent “machine”, the user must always be cautious as well as confirm the information received.

Share the news now

Source : Genk