Is ChatGPT Really a Game Changer or Is It a Race to the Bottom?
I first heard about ChatGPT (and AI overall) in the context of content creation. That was in December 2022, and social media influencers were touting it as the next big thing.
As a writer, this obviously concerned me. I was convinced that a career I had spent decades pursuing was all but over. After some sleepless nights, I signed up to see exactly what ChatGPT could do.
My First Experience With ChatGPT
At the time, I was working on an article on Fleetwood Mac’s greatest songs, so I asked ChatGPT what it knew about the song “Dreams.” Despite the text being grammatically correct, the information it gave was not. (Stevie Nicks wrote “Dreams,” not Christine McVie.)
The next night, I was able to sleep more soundly, assuming that the death of my career (and, in large part, the demise of human content creators) had been greatly exaggerated.
Yes, the cat is out of the bag — AI tools are not going anywhere. Various industries, such as customer service, human resources (HR), healthcare and, more specifically, content creation, have already started implementing them, but the application of ChatGPT and other AI-generated tools is still being defined.
So far, it seems like one big PR stunt put out by companies to appease investors who are jonesing to jump on the latest trend or a complete race to the bottom with misinformation no longer a concern as long as it generates clicks.
So, How Does ChatGPT Work?
ChatGPT was introduced to the world on Nov. 30, 2022, through its parent site, OpenAI, and was met with immediate and polarizing reactions, from blind panic to outright mockery. What wasn’t expected was how quickly it became popular. As of February 2023, Digital-Adoption.com ranked OpenAI as the 44th-most visited website in the world. Traffic has increased by over 3,500 percent, from 18.3 million upon to 672 million visits, since it first launched.
ChatGPT is a language model that is pre-trained on vast amounts of data. Once it has been trained, it can generate responses in real time, based on that data. As users interact with the tool and provide feedback on the responses generated by it, the model can be further trained and improved to become more accurate.
Artificial intelligence (AI) is transforming the workplace as I type, from automating routine tasks to providing insights that inform decision-making. But one of AI’s most promising (and most popular) applications is in the use of conversational AI. ChatGPT, which can understand and respond to natural language, leads the charge, but there are plenty of other apps out there from Microsoft’s Bing to Elon Musk’s allegedly soon-to-come TruthGPT.
Potential Applications of ChatGPT in the Workplace
Chatbots have been used by businesses for some time, and ChatGPT is an advanced chatbot that is already used in a variety of ways. For some industries, it is indeed promising. In customer service, it can handle routine inquiries, freeing up human agents to deal with more complex issues.
Human resources is another area where ChatGPT will prove popular, as it can automate the recruitment process, from writing compelling job descriptions to outreach emails to finding potential candidates and reducing the time and costs associated with scheduling interviews and assisting with basic HR-related questions. In healthcare, it can be used to automate appointment scheduling, prescription refills and triage. It may even provide patients with information about their conditions and treatments.
Use AI, but Don't Abuse It
Entrepreneurs are equally excited and concerned about the prospects of using AI in the workplace and its effects on employment. James Beattie, the founder of WeTested, a site that gives unbiased reviews on software products, told Work + Money that “businesses involving repetitive tasks or rely heavily on data processing are the most at risk of being automated by AI, like manufacturing, customer service, data entry and even some aspects of finance and accounting.”
But not all is lost for workers in these industries. Beattie believes that “workers and companies can prepare for the changes AI will bring by investing in continuous learning and adapting to new technologies.This might involve learning new skills, pursuing higher education or focusing on creative and strategic aspects of work that are harder to automate.
"AI may be able to augment human creativity and innovation in certain industries or roles, but it is unlikely to replace it entirely. Humans have a unique ability to think abstractly, empathize and make intuitive leaps that AI has not yet been able to replicate.”
Another entrepreneur, Ronaldo Hare, founder of Prem Property, states that jobs in writing, customer service and accounting “will be the first to completely fall, with human input only being needed to review the final result. Give the AI the right level of machine learning, and human input will not even be needed.”
Hare was proactive and began preparing for changes in his business. “As soon as we heard about ChatGPT, I got my team to research implementation into our property management company. [They] researched everything from ebook writing to rent calculation tools. It has affected us — and it's all we use to write the basis for our blogs, emails and much more. It reduces our admin and production hours, allowing us to focus on areas that are more profitable.”
He suggests businesses “embrace the change. Arrange a discussion with your team — first, to understand AI advancements and, second, to see how AI can benefit your business whatever industry or niche you are in [and finally] to create an implementation plan [and] ensure that it is moldable to the upgrades that will come. Always, remember to use AI, but do not abuse it.”
Journalism and Content Creation Is Where Things Get Murky
In April 2023, Business Insider’s global editor-in-chief Nicholas Carlson announced that the media company would start experimenting with the tool in a memo it published on its site. Carlson encouraged staff writers “to be excited about [ChatGPT’s] potential opportunities but not include any of its text in articles verbatim, as it can be untrue, plagiarized or simply be dull in tone."
A few months earlier, Buzzfeed stated it would begin using ChatGPT in quizzes and “other content,” and it’s been since discovered that at least 40 of the media sites' travel articles are largely AI generated with little to no human intervention. In other words, they all sound the same.
It should be noted that, at least in Buzzfeed’s case, shares rose 120 percent when the news of its use of ChatGPT was announced, giving more than a little credence to the idea that using AI instead of humans is something that is meant to please investors more than anything else. Then, about a month later, BuzzFeed shut down its news division for good and laid off about 180 people, or 15 percent of BuzzFeed's workforce.
Just the Facts
ChatGPT doesn’t always get things right (as we've mentioned) and, at this time, does not have a dedicated fact-checking feature. Therefore, it is up to the organization and individual to fact-check what ChatGPT provides. Remember, ChatGPT is designed to give responses based on the information it has been trained on, but that information is not always correct, especially if it’s been trained on biased or incomplete data.
Hopefully, in future iterations of ChatGPT, there will be a fact-checking feature that can help to verify the accuracy of the information provided. In the meantime, it's important for anyone using this tool to critically evaluate its responses and verify any information with trusted sources.
Plagiarism Is a Potential Problem
Since ChatGPT is a machine-learning model that generates responses based on already-existing data, it is possible for the tool to generate responses that are similar to existing content or text that has been published before. While ChatGPT is designed to generate responses based on patterns and correlations in data, rather than copying specific content from external sources, it can generate responses that are similar to existing content. But it does not intentionally plagiarize.
That said, it's still important to be mindful of plagiarism when using ChatGPT-generated content. Content creators should use a plagiarism checker (for example, Copyscape) to make sure that what is written doesn’t infringe upon another’s intellectual property. OpenAI does have its own plagiarism checker, but it's very new, and there are some kinks that have to be worked out.
Yas Ayub, the founder of SEO consultant agency Yaser.UK, agrees that plagiarism is a burgeoning issue. “A key component of SEO is creating new and optimizing pre-existing content from websites — mostly consisting of blogs, articles or rethinking whole pages. ChatGPT is a potentially dangerous tool in this field of the industry, as some content writers are exploiting the tool and getting it to write content that they then submit as their own work. This obviously presents enough of an issue in relation to copyright and stolen content.”
AI and Questionable Sources
Even more terrifying, some sources a journalist or content creator uses may act human but be entirely AI, and it will be incredibly difficult for them to tell the difference. In fact, this has already happened.
Julia Pugachevsky, a senior health reporter for Business Insider,put out a request on HARO (Help a Reporter Out) for sources who concealed their scars with tattoos. She was contacted by a breast-cancer survivor named Kimberly Shaw, who said that, through her tattoos, she was “reclaiming my body, taking back control of something that cancer had taken from me.”
After running Shaw’s statements through an AI text checker, Pugachevsky found she had been speaking with AI the entire time.
More Ethical Considerations of ChatGPT
While ChatGPT has the potential to revolutionize the workplace, there are also challenges that organizations face when implementing it. ChatGPT is allegedly trained on diverse data to avoid bias and discrimination. However, a study has shown that ChatGPT is given specific personas (as it was in an internalized setting), and it produced responses that wavered between toxic to overtly racist.
One of the biggest ethical considerations of ChatGPT is in avoiding bias and discrimination. If the data ChatGPT learns from is biased or otherwise discriminatory, the tool cannot discern right from wrong and will perpetuate that bias. Businesses must ensure that ChatGPT is trained on diverse data and that the output is regularly audited to avoid perpetuating any bias or discrimination.
According to Yazer.UK's Ayub, “When you take a deeper look into the way this AI technology was created, [you’ll find that] ChatGPT has had no access to the internet, or any new data or information, since it was created [in 2021]. This means the sources being used may be outdated, no longer relevant or have since been found to be incorrect. Researchers also would have chosen the sources used to educate the tool, so the technology may be subject to unconscious bias, either from the researchers and the coding and information they supplied it with or bias that may have been underlying within the sources or web pages it has access to.
“This could cause some serious issues if the tool was being used to find medical information or research, as it will likely not have the information you need — the medical world evolves very quickly, and there is no way the AI tool will be up to date. This could potentially be hugely dangerous if you use the tool to research how to respond to symptoms a patient may have or medication to prescribe to a certain illness. For example, you could receive outdated, disproved or just straight wrong responses, which ChatGPT will then argue with conviction.”
Privacy and Data Concerns of ChatGPT
Privacy and data security is another issue. Governments and businesses must ensure that the data collected and processed by ChatGPT is protected and users’ privacy rights are respected. Currently, Italy has banned ChatGPT as the result of a possible data breach of Europe’s privacy regulations.
An Italian privacy regulator, Garante, found that users could view the titles of conversations other users were having with the tool. The regulator is also concerned about the lack of age restrictions for ChatGPT and — as noted earlier — its many inaccuracies and biases.
The EU is now considering the European AI Act, which will heavily restrict the use of all AI (including ChatGPT) in education, infrastructure, law enforcement and the judicial system. The U.S. currently has no regulations or restrictions on the use of AI. However, the nonprofit research group, the Center for AI and Digital Policy, submitted a complaint to the Federal Trade Commission that the language model is “biased, deceptive and a risk to privacy and public safety” — and is in violation of its AI guidelines.
ChatGPT's Environmental Impact
One more thing: ChatGPT uses a lot of energy. In January 2023 alone, the tool used 1.287-gigawatt hours, which is what approximately 120 homes in the United States use in a year’s time.
So, as we continue to show more concern for environmental issues and face stricter government regulations, the use of AI could be seen as simply irresponsible from an environmental standpoint.
What Does All of This Mean for Human Content Creators?
We won’t kid you — regardless of these issues, some people will likely lose their jobs. But there will also be a growing demand for those who specialize in high-quality data to ensure that ChatGPT is accurately trained to produce relevant responses. Content creators may also be used to craft responses to common queries and conversations and for creating content ensuring ChatGPT aligns with a brand’s voice and tone.
This means the content creator will edit ChatGPT’s responses to have more humanity and accuracy. As the tool becomes more integrated with other AI technologies, it may also begin to develop multimedia content.
Music journalist and educator Lily Moayeri said, "I would love to see AI take over more of the uninspiring parts of the job, like regurgitating press releases as news items [and], in the process, freeing up time and energy for more meaningful articles that require my personal experience, thoughts, opinions, interpretation and perspective, which is something AI will never be capable of generating on its own."
The Future Is Now
ChatGPT is here to stay and, in some cases, will be a powerful tool for business. However, while it has the potential to revolutionize the workplace, there are challenges and ethical considerations that have not yet been addressed and may not be for a long time, particularly in journalism and content creation.
As this is the case, I think it’s far too early to remove human content creators from the equation, and as one myself, it takes as long for me to edit a piece for tone, grammar and accuracy as it does to write it from scratch. Nevertheless, I will continue to use it for outlines, brainstorming and to bust through any writer’s block that may crop up.
I checked the tool again a few weeks later to ask about Fleetwood Mac’s “Dreams.” This time, the answer it gave was more accurate, but it read more like a Wikipedia entry — devoid, obviously, of any emotion or human tone.
Needless to say, I still sleep well at night.