Four months ago, a small company in San Francisco took the technology world by storm by launching a new online chatbot, ChatGPT. It can answer complex questions, write poems and even mimic human emotions.
According to the New York Times, OpenAI has 375 employees. However, the startup company has received billions of dollars of investment from many other investors including Microsoft. The company said on Tuesday that its new technology is designed to underpin the GPT-4 engine. It has powered everything from search engines to personal online tutors to chatbots and other systems. Most people will be able to use this technology through the new version of the company’s ChatGPT chatbot.
In just a few months, OpenAI’s breakthrough has given the tech industry its biggest shakeup in decades, according to industry experts. It’s a fundamental technological change as important as the development of web browsers in the early 1990s. Its incredible rapid development has surprised computer scientists.
GPT-4 is capable of analyzing large amounts of data collected from the Internet. For example, it is capable of providing answers to the quality of the top 10 candidates who passed the US Uniform Bar Exam. ChatGPT’s answers were similar to those of the worst-performing 10 percent of examinees. It can instantly provide one’s tap count and photo details. Al Jazeera reported that it could show “human-level” success in postgraduate tests at US universities.
GTP-4 is the product of intensive study. His storehouse of knowledge is almost close to man. The company said in a blog post, “We spent six months of research building GPT-4, taking lessons from our previous pilot program as well as ChatGPT.” This resulted in our best ever results in terms of reliability, manageability and accuracy. But it is still far from perfect.
In an online demonstration, OpenAI President Greg Brockman demonstrated that GPT-4 was able to create a real website based on hand-drawn images.
OpenAI says the new chatbot can beat humans 90 percent of the time on the evidence-based reading and writing and speaking sections of the GRE, or Graduate Record Examination, Scholastic Assessment Test. ChatGPT didn’t score that high. GPT-4 scored 1300 marks out of 1600 in SAT exam. The GPT-4 can score 5 out of 5 in biology, calculus, macroeconomics, psychology, statistics, and history on the high school Advanced Placement Test.
According to the company, GPT-4 is also less likely to give wrong, offensive and insulting answers than ChatGPT. It specializes in various subjects. Impressed even doctors with medical advice. When provided with an article from the New York Times, the new chatbot can provide an almost concise and accurate summary. If you add a random sentence to the summary and ask the bot if the summary is wrong, it will show the added sentence.
New chatbots will threaten many professions, from journalism to healthcare. Tech giants such as Google, Microsoft, Huawei, Alibaba and Baidu have jumped into competition in the fast-growing artificial intelligence sector. But the New York Times says that OpenAI is not equal to human intelligence. It is often unable to understand what is true and what is false. So may give completely wrong information.