(CNN) The first day after the reveal GPT-4 amazed many users in early tests and a company demo with its ability to design suits, pass standardized exams, and create a working website from a hand-drawn sketch.
On Tuesday, OpenAI announced the next generation of artificial intelligence technology powering its viral chatbot tool ChatGPT. The more powerful GPT-4 promises to blow up previous iterations and potentially transform the way we use the internet to work, play and create. But it could also contribute to challenging questions about how AI tools can disrupt professions, enable students to cheat, and transform our relationship with technology.
GPT-4 is an updated version of the company’s large language model, which is trained on vast amounts of online data to generate complex ones Responses to user prompts. It’s now available via a waitlist and has already found its way into some third-party products, including Microsoft’s new AI-powered Bing search Engine. Some users with early access to the tool share their experiences and highlight some of its most compelling use cases.
Here’s a closer look at GPT-4’s potential:
Analyze more than just text
At its core, the biggest change to GPT-4 is the ability to work with photos that users upload.
One of the most mind-blowing use cases so far came from an OpenAI video demo showing how a Drawing could be turned into a functional website In minutes. The demonstrator uploaded the image in GPT-4 and then pasted the resulting code into a preview showing what a working website might look like.
In its announcement, OpenAI also showed how GPT-4 was asked to explain a joke from a series of images – showing a smartphone with the wrong charger – and described why it was funny. While it may sound easy, the dissection is a joke Because of the context they need, they are more complicated for artificial intelligence tools to capture.
In another test, The New York Times GPT-4 showed an image of the inside of a refrigerator and asked it to create a meal based on the ingredients.
The photo feature isn’t live yet, but OpenAI is expected to roll it out in the coming weeks.
Coding made even easier
Some early GPT-4 users with very little to no programming knowledge also used it Recreate iconic games like Pong, Tetris or Snake after following the step-by-step instructions provided by the tool. Others have made their own original games. (GPT-4 can write code in all major programming languages, according to OpenAI.)
“The powerful language capabilities of GPT-4 are used for everything from storyboarding to character creation to game content creation,” said Arun Chandrasekaran, analyst at Gartner Research. “This could lead to more independent gaming providers in the future. But beyond the game itself, GPT-4 and similar models can be used to create marketing content around game previews, generate news articles, and even moderate gaming discussion forums.”
Similar to gaming, GPT-4 could change the way people develop apps. A user on Twitter said so created a simple drawing app in minutes while another claimed to have it coded an app that recommends five new movies every day, along with providing trailers and details on where to watch them.
“Coding is like learning to drive a car — as long as the beginner gets some guidance, anyone can code,” said Lian Jye Su, analyst at ABI Research. “AI can be a good teacher.”
Exams passed with flying colours
Although OpenAI said the update is “less powerful” than humans in many real-world scenarios, it shows “human-level performance” across various professional and academic tests. The company said that GPT-4 recently passed a mock bar exam at a law school with a score of about the top 10% of test takers. In contrast, the previous version GPT-3.5 only reached the bottom 10%. The latest version According to OpenAI, it also performed strongly on the LSAT, GRE, SATs, and many AP exams.
In January, ChatGPT made headlines for its ability to pass prestigious graduate-level exams, such as one from the University of Pennsylvania’s Wharton School of Business, but not with particularly good grades. The company said it spent months learning lessons from its testing program and ChatGPT to improve the system’s accuracy and ability to stay on topic.
Give more specific answers
Compared to the previous version, GPT-4 is able to provide longer, more detailed and more reliable written answers, according to the company.
The latest version can now provide answers of up to 25,000 words, up from around 4,000 previously, and can provide detailed instructions for even the most unique scenarios, ranging from cleaning a piranha’s aquarium to extracting DNA from a strawberry. An early user said it provided detailed pickup line suggestions based on a question listed on a dating profile.
Optimizing work in different industries
Joshua Browder, CEO of legal services chatbot DoNotPay, said his company already is work on using the tool Generating “one-click lawsuits” to sue robocallers, in an early indication of GPT-4’s tremendous potential to transform the way people work across industries.
“Imagine receiving a call, clicking a button, [the] The call is transcribed and a 1,000-word lawsuit is generated. GPT-3.5 wasn’t good enough, but GPT-4 does the job extremely well,” Browder tweeted.
Meanwhile, Jake Kozloski, CEO of dating site Keeper, said his company is with the tool too better suit its users.
According to ABI Research’s Su, it’s possible we’ll see big advances in the “connected car” as well [dashboards]Remote healthcare diagnostics and other AI applications that were previously not possible.”
In progress
Although the company has significantly improved its AI model, GPT-4 has limitations similar to previous versions. OpenAI said the technology has no knowledge of events that occurred before his dataset was truncated (September 2021) and does not learn from his experiences. It can also make “simple reasoning mistakes” or be “too gullible in accepting obvious false statements from a user” and not double-check work, the company said.
Gartner’s Chandrasekaran said this also reflects many of today’s AI models. “Let’s not forget that these AI models aren’t perfect,” said Chandrasekaran. “They may produce inaccurate information from time to time and be black box in nature.”
For now, OpenAI said GPT-4 users should exercise caution and exercise “great diligence,” especially “in high-stakes contexts.”