Scary Good: 
A Human discussion About ChatGPT

Artificial intelligence is infiltrating our lives in ways we realize and don’t realize. Is it hype or hope?

Scary Good: A Human Discussion about ChatGPT - Douglas Magazine June/July 2023

In November 2022, San Francisco-based OpenAI released its first version of ChatGPT, which by the following April had morphed into ChatGPT4. Within five days of its launch, it drew more than one million users and by January became the fastest-growing app of all time, with 100 million active users. 

Once you’ve created an account and downloaded the app, it’s disarmingly easy to use. It can generate, edit and interact with users employing both creative and technical writing. It can compose a song, write a screenplay, or teach a course. Just type in a statement or request, and follow the system’s various prompts.

But as something of a rhetorical question, does a business need ChatGPT (an acronym for Chat Generative Pre-Trained Transformer)? Businesses have already been using AI for years by doing a Google or Bing search, asking Alexis or Siri a question or interacting with an online assistant. One Victoria brewery that used ChatGPT to write product descriptions and social media posts went further, using it to formulate a recipe for a pale ale. And ChatGPT’s value in medical research and health care cannot be ignored.

Yet, this wildly popular app is not without flaws. ChatGPT itself admits it can give wrong answers: “Yes, language models can generate incorrect answers,” says the app. “However, they can be trained to become more accurate over time.”

The Human Touch

ChatGPT uses RLHF (reinforcement learning from human feedback, so ergo, it’s only as good as the guidance it gets via artificial conversations. Crucial to its correctness is the ability, and even world view, of the human trainer: “The accuracy of a language model depends on the quality of the data it is trained on and the skill of the person who is training it.”

OpenAI is already facing defamation lawsuits by those whose reputations have been sullied due to incorrect information. Others are concerned about privacy. And stories of university cheaters and criminal acts, particularly cybercrime, are flourishing.

ChatGPT is not limited to answering questions, such as, “Who is Premier David Eby?,” which, by the way, supplied an incorrect answer because training data stopped in 2021. Its greater strength lies in its ability to write code that can generate scripts and functions based on instructions. It can supply detailed explanations for complex subjects; it can solve thorny mathematical problems; it can compose musical lyrics; it can write text and even poetry in different styles.

Victoria Realtor Jane Johnston used an artificial intelligence program to generate copy for a golf course listing and amidst the copy was a factual mistake. Johnston, with RE/MAX Camosun, also tested AI by using it to introduce new partners. “I found it a bit generic, lacking the human touch,” she says. “It’s not particularly original.”

The Call for a Moratorium

Johnston notes that if someone is having writer’s block and cannot put words to a page, AI-generated copy provides a great starting point. “But I wouldn’t rely on it,” she says.

In a December 2022 tweet, Elon Musk, who was one of the founders of OpenAI, wrote, “ChatGPT is scary good. We are not far from dangerously strong AI.” But by late March 2023, Musk, Apple co-founder Steve Wozniak and hundreds of other tech experts called for a stop to the training of AI systems for at least six months due to “profound risks to society and humanity.”

On the flip side, Bill Gates believes ChatGPT can provide workers with more free time by making them more efficient, teaching high-level math and providing medical guidance. It can make the work of white-collar and health care employees easier.

At the University of Victoria, artificial intelligence expert Homayoun Najjaran likens AI, such as ChatGPT, as the ongoing process to create machines that  will make jobs easier. “Some people may become irrelevant. Some may become more relevant,” says Najjaran, a professor of mechanical engineering. “Maybe in 10 years, I will be irrelevant. So, I have to change the way I teach.”

He admits apps like ChatGPT will likely eliminate some jobs. Research indicates white-collar work is most at risk, including tasks done by lawyers and computer programmers. So, knowing how to use ChatGPT could be a job saver. 

Inventions like ChatGPT are not the first tools that generate answers and lighten the load. There was the abacus, then the slide rule, then handheld calculators, then desk computers, then smartphones, all geared to increasing efficiency and, in some cases, freeing people from doing dangerous or menial work, Najjaran says. “Go back 200 years, where a huge building project may have had one engineer and 20,000 workers. Today, it would be 200 engineers, 200 workers and 3,000 machines.”

And if ChatGPT can allow someone doing a dead-end job to learn new skills that enables them to get a better-paying job, is that not good for society? If efficiency can be increased tenfold or if research can be done faster, is that not beneficial?

Yet, from the waterwheel to Jacquard’s weaving loom to an assembly-line robot, inventions that can be programmed still need a human behind them. “There is no connection between AI and nature. We need a bridge. The first touch has to come from humans, the human connection,” Najjaran says.

As for the chance ChatGPT and its ilk will become like HAL 9000 in 2001: A Space Odyssey, Najjaran believes, “Good or evil, it’s in our hands. You cannot say, stop building tools because it’s dangerous. A hammer is a tool, but it too can be dangerous.” 

A Robot Wrote This
We asked ChatGPT to weigh in on the question of job security for humans.

Artificial intelligence (AI) has the potential to automate certain tasks and make some jobs obsolete. However, it’s important to note that AI is not a replacement for human workers. Instead, it can complement and enhance their work, freeing up time for more valuable tasks and improving productivity.

“While some jobs may become automated in the future, new jobs will also emerge as a result of technological advancements. AI requires humans to develop, maintain, and program it, which means there will be a need for people with the necessary skills and knowledge to work alongside it.

“It’s important to adapt to changes in the job market by learning new skills and staying up-to-date with the latest technological developments. By doing so, individuals can ensure they remain relevant and valuable in the workforce, regardless of changes brought about by AI.”

Shannon Moneo is a graduate of the University of Regina’s journalism program and has been writing since 2001. After living in 16 cities, towns and villages in five provinces, she has made Sooke her home since 1992.