The company behind the ChatGPT chatbot has rolled out its latest artificial intelligence model, GPT-4, in the next step for a technology that鈥檚 caught the world鈥檚 attention.
The new system can figure out tax deductions and answer questions like a Shakespearan pirate, for example, but it still 鈥渉allucinates鈥� facts and makes reasoning errors.
Here鈥檚 a look at San Francisco-based startup OpenAI鈥檚 latest improvement on the generative AI models that can spit out readable text and unique images:
WHAT鈥橲 NEW?
OpenAI says GPT-4 鈥渆xhibits human-level performance.鈥� It鈥檚 much more reliable, creative and can handle 鈥渕ore nuanced instructions鈥� than its predecessor system, GPT-3.5, which ChatGPT was built on, OpenAI said in its announcement.
In an online demo Tuesday, OpenAI President Greg Brockman ran through some scenarios that showed off GPT-4鈥檚 capabilities that appeared to show it鈥檚 a radical improvement on previous versions.
He demonstrated how the system could quickly come up with the proper income tax deduction after being fed reams of tax code 鈥� something he couldn鈥檛 figure himself.
鈥淚t鈥檚 not perfect, but neither are you. And together it鈥檚 this amplifying tool that lets you just reach new heights,鈥� Brockman said.
WHY DOES IT MATTER?
Generative AI technology like GPT-4 could be the future of the internet, at least according to Microsoft, which has invested at least $1 billion in OpenAI and made a splash by integrating AI chatbot tech into its Bing browser.
It鈥檚 part of a new generation of machine-learning systems that can converse, generate readable text on demand and produce novel images and video based on what they鈥檝e learned from a vast database of digital books and online text.
These new AI breakthroughs have the potential to transform the internet search business long dominated by Google, which is trying to catch up with its own AI chatbot, and numerous professions.
鈥淲ith GPT-4, we are one step closer to life imitating art,鈥� said Mirella Lapata, professor of natural language processing at the University of Edinburgh. She referred to the TV show 鈥淏lack Mirror,鈥� which focuses on the dark side of technology.
鈥淗umans are not fooled by the AI in `Black Mirror鈥� but they tolerate it,鈥� Lapata said. 鈥淟ikewise, GPT-4 is not perfect, but paves the way for AI being used as a commodity tool on a daily basis.鈥�
WHAT EXACTLY ARE THE IMPROVEMENTS?
GPT-4 is a 鈥渓arge multimodal model,鈥� which means it can be fed both text and images that it uses to come up with answers.
In one example posted on OpenAI鈥檚 website, GPT-4 is asked, 鈥淲hat is unusual about this image?鈥� It鈥檚 answer: 鈥淭he unusual thing about this image is that a man is ironing clothes on an ironing board attached to the roof of a moving taxi.鈥�
GPT-4 is also 鈥渟teerable,鈥� which means that instead of getting an answer in ChatGPT鈥檚 鈥渃lassic鈥� fixed tone and verbosity, users can customize it by asking for responses in the style of a Shakespearean pirate, for instance.
In his demo, Brockman asked both GPT-3.5 and GPT-4 to summarize in one sentence an article explaining the difference between the two systems. The catch was that every word had to start with the letter G.
GPT-3.5 didn鈥檛 even try, spitting out a normal sentence. The newer version swiftly responded: 鈥淕PT-4 generates groundbreaking, grandiose gains, greatly galvanizing generalized AI goals.鈥�
HOW WELL DOES IT WORK?
ChatGPT can write silly poems and songs or quickly explain just about anything found on the internet. It also gained notoriety for results that could be way off, such as confidently providing a detailed but false account of the Super Bowl game days before it took place, or even being disparaging to users.
OpenAI acknowledged that GPT-4 still has limitations and warned users to be careful. GPT-4 is 鈥渟till not fully reliable鈥� because it 鈥渉allucinates鈥� facts and makes reasoning errors, it said.
鈥淕reat care should be taken when using language model outputs, particularly in high-stakes contexts,鈥� the company said, though it added that hallucinations have been sharply reduced.
Experts also advised caution.
鈥淲e should remember that language models such as GPT-4 do not think in a human-like way, and we should not be misled by their fluency with language,鈥� said Nello Cristianini, professor of artificial intelligence at the University of Bath.
Another problem is that GPT-4 does not know much about anything that happened after September 2021, because that was the cutoff date for the data it was trained on.
ARE THERE SAFEGUARDS?
OpenAI says GPT-4鈥檚 improved capabilities 鈥渓ead to new risk surfaces鈥� so it has improved safety by training it to refuse requests for sensitive or 鈥渄isallowed鈥� information.
It鈥檚 less likely to answer questions on, for example, how to build a bomb or buy cheap cigarettes.
Still, OpenAI cautions that while 鈥渆liciting bad behavior鈥� from GPT is harder, 鈥渄oing so is still possible.鈥�
By Kelvin Chan
THE ASSOCIATED PRESS
