ChatGPT is changing everything. But it still has its limits

The AI chatbot has proven to be very capable of technical tasks, such as writing and coding. But it can't do everything - yet.
Written by Sabrina Ortiz, Editor
chatgpt illustartion on a futuristic background
Image: Getty Images / NurPhoto

Since its release in late November, ChatGPT has taken the world by storm. The chatbot's advanced AI abilities allow it to do tasks completely on its own, such as composing essays, emails and poems, writing and debugging code, and even passing exams. Now that a chatbot can do what humans do so well in a matter of seconds, what does that mean for our future?

If you have had the chance to chat with the AI chatbot, you were probably impressed with how much it can understand and its ability to respond in a conversational manner. However, the chatbot is capable of doing much more, and its technical capabilities are tested every day. 

Also: What is ChatGPT and why does it matter? Here's everything you need to know

Already, ChatGPT is changing expectations and practices in education, as well as in professional fields that involve skills like coding. While some are sounding the alarm bells, and others are taking the advancements in their stride, technology analysts are still assessing how the tool will change things. 

For instance, a professor at the University of Pennsylvania's Wharton School of Business used ChatGPT to take an MBA exam where it passed with a score of B to B-, leaving the professor impressed with its explanations and operations management. 

Also: The best AI art generators: DALL-E 2 and alternatives

Professors at Minnesota University Law School gave ChatGPT four separate law school final exams and then graded the exams blindly, shuffling it with actual student exams. The bot passed with an average grade of C+, according to a whitepaper. Although this score is on the lower end of what students score, it is impressive for a bot. 

"We expect such language models to be important tools for practicing lawyers going forward; we also expect them to be very helpful to students using them (licitly or illicitly) on law school exams," said the Minnesota University Law School professors in the whitepaper about the experiment.  

As ChatGPT's capabilities continue to be tested, one of people's biggest concerns about ChatGPT is how it will affect the education system. Will its technical proficiency make learning certain skills for humans obsolete? Will people learn less in school because of the temptation to use ChatGPT to do the work for them?

Schools across the country are already taking measures to try to curb such cheating and slacking in the classroom. The New York City Department of Education, for instance, blocked student and teacher access to ChatGPT on its networks. 

The issue has also made its way to higher education, with professors including rules for using ChatGPT in their syllabuses. How professors are handling the AI depends on each individual, with some banning it entirely and others taking a fairly open approach. 

An associate professor at Wharton, Ethan Mollick, shared on LinkedIn his syllabus's policy for ChatGPT, which allows for its integration into students' work. "We don't have many answers yet [about AI], but we should welcome discussion. And teach how to use the tools responsibly," said Mollick.  

Just another tool in the toolbox?

Embracing AI generative technologies could actually help enrich students' education, according to Sid Nag, cloud services and technologies analyst at Gartner. 

"It's like saying, 'Is the use of a calculator going to hinder the quality of people's ability to add one and one plus two?' No."

A factor to remember when discussing AI is that a lot of the tasks it is capable of doing are tedious and time-consuming in nature. In many instances, doing these tasks doesn't add much value to the overall goal. 

Also: How to get started using ChatGPT

For example, if you are doing a major research project, writing an introduction to the paper or writing code is a very small portion of the project, with thorough analysis and interpretation of that code being at the forefront. 

Over time, the technologies underlying ChatGPT will "dramatically" automate many of the mundane writing and coding tasks  Rajeesh Kandaswamy, AI innovation analyst at Gartner, tells ZDNET.  

"But, there is a lot of creative and unique writing and coding that involves imagination, synthesis and other complexities that won't be easily solved shortly by these AI tools," Kandaswamy points out.

In addition to the education system, people often worry about generative AI tools taking over human jobs. However, as advanced as these models are, they still require a human to produce the correct output. AI chatbots are very good at delivering words or code - but less good at knowing whether it is any good, or even accurate. Without human direction from a person well-versed on the subject at hand, these tools are much less useful. 

"Just because you have a toolbox doesn't mean you know how to fix a broken appliance, right?" Nag says. 

A new category of jobs 

As a matter of fact, the more that these models grow, the more they will require assistance -- creating a need for new jobs and even more studies. If some menial job tasks are eliminated, there's a chance those roles will simply shift into other ones that are created to assist AI.

ChatGPT and its underlying capability will have a lot of impact on some of these skills says Kandaswamy. "But these capabilities can also lead to new products and services either directly through what these technologies do or in building upon what they do."

Also: ChatGPT can write code. Now researchers say it's good at fixing bugs too

Lastly, before we see any major transformation in our everyday life, the technology has to be reliable and proven to make no errors, and right now, we are not at that point -- yet.  

"ChatGPT is prone to mistakes and generating wrong output (hallucinations). That does not make it useless, but it does make it unreliable, for now," Kandaswamy continues. "We will require supporting mechanisms, tests, and other quality checks to use it at a scale where quality and accuracy are important."

The future of generative AI sources such as ChatGPT are promising and have the potential to significantly change the way we do things. Right now, we might be at the same point we were decades ago when the internet was first introduced: lots of potential but still a long way to go. And while tools like ChatGPT are opening up new possibilities, they are far from perfect.

"Yes, this is revolutionary, and I won't be surprised if it surpasses the internet in impact across business and society. Only time will tell. We are in the very early days," says Kandaswamy. 


ZDNet's Monday Opener is our opening take on the week in tech, written by members of our editorial team. 


Editorial standards