ChatGPT, an artificial intelligence chatbot that has become a well-known tool, enabling ideas, information and writing to be automatically generated by anyone with the hit of a button, grew to prominence in November 2022. As per a Forbes article on the subject, ChatGPT rapidly gained traction on social media. In as little as five days, the artificial intelligence tool had attracted over a million curious users.
UW has taken note of this technological advancement and the widespread popularity of its use. There is no denying that a system manufactured to spit out the answers or even the essay you ask it to write is not without ethical concerns and controversy. Having always enjoyed the written word, I found myself questioning whether the value of writing would be lost in a world where auto-generated pieces of writing is the new normal.
In search of answers, I reached out to Anindya Sen, professor of economics and acting executive director of the Waterloo Cybersecurity & Privacy Institute (CPI) at UW, to discuss the rise, direction, and role of ChatGPT and the launch of a new course on LEARN, titled ‘AI and Society’, focused on teaching AI literacy skills for faculty, students, and staff. The course is not currently planned to count for course credit, however, it is now accessible under ‘Self-Registration’ on Learn and is a self-paced online course.
Considering the role of AI: ready to replace human workers?
Sen expressed that while the quick evolution and accuracy of AI is remarkable, he feels that for many professions, AI presently serves to “enhance” rather than replace human workers. He offered a great example by calling to mind the work of radiologists, who work to detect anomalies in imaging results. When using AI in radiology, Sen expressed, “These machine learning models, they’re very accurate and [can detect] anomalies the human eye [cannot].”
AI enables radiologists to become more productive, but AI “won’t immediately replace radiologists.” Human workers would remain essential in reviewing what AI has detected to ensure accuracy, particularly in healthcare fields in which correct diagnosis is paramount to the profession. In his teaching, Sen believes in having students use AI responsibly to enhance their productivity. He mentioned that in the past, teachers were reluctant to allow the use of calculators, however in viewing how it enables students to make faster calculations, it has become an accepted tool.
Considering the ethical issues and caveats of AI tools
Unlike calculators, Sen expressed that the use of ChatGPT brings to mind many concerns, including data collection, how user data is stored, how individual privacy is impacted, and data scraping. Concerns regarding cheating and plagiarism on written assignments is another key ethical dilemma surrounding use of AI. An online guide on generative artificial intelligence use at UW, by the Copyright Advisory Committee and Dan Brown, a Professor from the Cheriton School of Computer Science, outlines how GenAI is virtually impossible to attribute accurately as well as concerns regarding lack of clarity of copyright ownership of GenAI content in Canada.
Additionally, Sen described how “news organizations in Canada have launched a case against OpenAI with allegations that they’re unfairly copying and scraping all their content to feed their algorithm.” The scraping of copyrighted data is unlawful and problematic, in the case of OpenAI, it was accused of taking data without compensating the companies it had taken this same data from. He went on to describe how The New York Times also opened a lawsuit against OpenAI for data scraping. He concludes that AI introduces concerns regarding intellectual property infringement. This also brings to mind conversations around what kind of data can be considered public domain and under which circumstances data can be taken. Overall, Sen states that while “there’s nothing wrong in building a model based on these news articles, but [it requires] acknowledgment [of the source] and if there are other agencies who have already collected this information, they should be compensated.” These are key ethical concerns he strongly insists require answers and consideration as we move forward into a world where AI and ChatGPT continue to expand.
Looking towards the future: UW’s first AI course
Sen believes that the new AI course can help all students learn something new. For students not enrolled in math and engineering who may be less familiar with certain AI tools, he believes that the definitions in the course that describe AI, machine learning, and algorithms can help students gain a better understanding of such technologies. The course itself also dives into the unique ethical and social concerns surrounding AI. The course has been designed to enable thoughtful use and understanding of a tool that is sure to continue to expand its ability and usage. Sen shared that he will be speaking with faculty deans at UW this January 2025, “to see how [they] can make this course available to all students and [make it] something they can take for course credit.”
Where AI falls short: a glimpse into the power of human emotion and experience
Sen highlights the benefits and implications of AI use, however, when it comes specifically to the field of writing, another prominent figure at UW has done research into the realm of music writing and what makes certain lyrics and songs so meaningful to listeners. Paul Thagard, cognitive scientist, full-time writer, and distinguished professor emeritus of philosophy at UW, looked into why so many find comfort in sad songs. Taylor Swift is well-known for her heartbreak ballads and for many who find solace in her tunes, it is the relatability and the authenticity in her lyrics that make fans feel understood. Thagard emphasizes, “What makes the sad songs so powerful for us is that they engage our emotions.” Given sadness is a universal human emotion we experience often enough, music can enable us to work through such emotions. In a chapter of his new book, Dreams, Jokes and Songs: How the Brain Builds Consciousness, Thagard discusses Swift’s hit song, All Too Well, and the powerful range of emotions that are expressed in the ballad, as being key to its fan-favourite status. Thagard’s work and research highlights that AI can write lyrics and poetry but fails to authentically replicate the human capability in one area that makes a song truly meaningful: emotions.
ChatGPT is a tool and as our world continually advances in the realm of technology, there are bound to be emerging technologies that bring up ethical concerns and controversial uses. Technology is an amazing part of our lives, enabling and advancing society to incredible heights. It is key that even as our world progresses and revolutionizes itself, we remember that human communication, which is thoughtful, present, and honest, remains more impactful than generic, inputted, and reused writing ever will be. There will never be a voice exactly like yours. Instead of asking ChatGPT to write your essay in its voice, perhaps consider writing it from your perspective: human, real, and encompassing truths, experiences, and emotions that AI will never match.
