Menu Menu
[gtranslate]

ChatGPT-4 is OpenAI’s most powerful language tool yet

There is no slowing the meteoric rise of generative AI, and OpenAI has just rolled out its most powerful model yet, ChatGPT-4. What’s different with the latest iteration? 

Of the many Silicon Valley giants vying to own the generative AI space, Microsoft has struck gold with OpenAI’s ChatGPT. 

Following its full-scale integration with company-owned search engine Bing, OpenAI has unveiled the fourth iteration of its AI language model, ChatGPT-4, which is attracting huge commercial interest already. 

Touted as being ‘more creative and collaborative than ever before,’ the tech has already been picked up by several companies’ development teams.

Morgan Stanley has put it to organising its wealth management data, for instance, while Stripe Inc is autonomously refining its combat fraud barriers. Even foreign language-learning app Duolingo is using ChatGPT-4 to help users practice real-world conversations and explain mistakes.

These are but a few early examples of how the generative language AI is disrupting global job markets, but just how much better is ChatGPT-4 than its predecessor? 

The distinction has been described by a company announcement as ‘subtle,’ with the key upgrade being its new multi-modal feature. What this means, is that the new system can accept both text and image inputs – unlike OpenAI’s image generator DALL-E, however, the latter stimulus will be interpreted or solved with the response coming back purely in text. 

Whether the user wants a visual mathematics problem solved, or the nuance of an abstract meme explained, ChatGPT-4 navigates the prompts brilliantly. See the below example, and how eerily human-like the clarification is.

Still able to complete all its previous recreational feats; whether that’s writing scripts, social media copy, or cooking recipes for everything left in your fridge, a number of new benchmarks have underlined its performance enhancements. 

Purely as a showcase of the technology, the company revealed that ChatGPT-4 aced lawyers’ Uniform Bar Exam, LSAT SAT Maths, and SAT Evidence-Based Reading & Writing exams, scoring in the 88th percentile and above.  

Given various GPT models have been banned across several US States to stop students cheating assignments, demoing GPT-4 with various academic tests was certainly an odd decision. Nevertheless, it pulled through with flying colours. Here are the full results.

When talking about its key function: providing nuanced, curated responses asked on the internet, the newest model still has some limitations. For one, the dataset which trained the system was cut off at 2021, meaning prompts regarding new subject matter are hit and miss depending on what was public record at the time. 

On the upside, though, users are less likely to break Bing’s guardrails and prompt the bot to share dangerous advice, threatening sentiments, or straight-up misinformation. There are still instances of so-called ‘hallucinations’ and unhinged behaviour, but theoretically, they should be few and far between now. 

Owing to six months of safety training and tweaking, it is now reportedly ‘82 percent less likely to respond to requests for disallowed content and 40 percent more likely to produce factual responses than GPT-3.5,’ an OpenAI blog post says. 

Aside from toying around with the new Bing, getting full access to ChatGPT-4 and its exciting new image function requires signing up to ChatGPT plus for $20 a month. Whether or not the upgrade merits that outlay, will become clearer in the near future. 

Accessibility