Artificial intelligence company %OpenAI has released a more powerful chatbot called “GPT-4.”
The San Francisco-based company said that GPT-4 represents a new version of its wildly popular artificial intelligence chatbot %ChatGPT.
According to privately held OpenAI, the new GPT-4 model can respond to images and provide recipe suggestions from photos of various ingredients. It can also identify and interpret photo captions and drawings.
GPT-4 is capable of processing up to 25,000 words, about eight times more than its predecessor ChatGPT.
OpenAI said that its scientists and engineers spent six months on safety features for GPT-4 and have trained it to respond appropriately to human feedback.
However, the company warned that its artificial intelligence technology is still evolving, and that GPT-4 could be prone to sharing disinformation of getting search inquiries wrong.
GPT-4 will initially be available to ChatGPT Plus subscribers, who pay $20 U.S. per month for premium access to the chatbot.
GPT-4 is known as generative artificial intelligence. It uses algorithms and predictive text to create content based on prompts from humans.
OpenAI said in releasing GPT-4 that it has more advanced reasoning skills than ChatGPT. In a live demo, for example, GPT-4 generated the correct answer to a complicated tax question.
ChatGPT has become the most popular app in history since it launched in November 2022, with more than 100 million people downloading it in less than six months.
People have used ChatGPT to write songs, poems, marketing copy, emails, computer code, and to help complete homework assignments such as essays.
ChatGPT is now powering the %Microsoft ($MSFT) %BingSearchEngine after the tech giant invested $10 billion U.S. in OpenAI.
Despite all the success, there are concerns that chatbots will take human jobs. Chatbots such as ChatGPT are also known to “hallucinate,” as engineers call it, where they invent facts or makes reasoning and logic errors.