Been Hearing About ChatGPT? Here’s What You Need to Know
By: Michael McQueen
If you have been anywhere close to LinkedIn or business news over the past few weeks you would have encountered the conversation surrounding ChatGPT.
Launched in November 2022 by OpenAI and stirring up the tech world since, this new AI chatbot is unparalleled in its capacity to seem truly human. Able to carry out flawlessly natural conversations, as well as replicate the language of humans across a range of contexts, this technological innovation and its many applications are raising some serious questions.
Behind the chatbot is a network of artificial neurons which replicates the human brain and has analysed endless information resources on the internet. [1] This technology is able to imitate many of what we consider the most human of our language processes: constructing an argument, telling a joke, simplifying jargon and brainstorming. [2]
Much of the attention recently stirred up surrounding it was a result of university educators becoming aware of its capacity to evade plagiarism detection tools specifically designed to detect AI-generated text. Deakin University professor claimed to uncover the use of the software in one fifth of her students’ essays, exploited for its ability to construct essays and arguments which are indistinguishable from those developed by human students. Beyond simply endangering student learning, some of the exams that ChatGPT has been revealed to be capable of passing are medical licensing exams, the veracity of which is literally of life-or-death importance.
Such technology raises obvious problems for educators in their marking, prompting a rethink of the current framework of student assessment.[3] Established plagiarism detection organisations such as Turnitin are now facing the challenge of developing their software to account for the new and improved plagiarism strategies of students. In response to such issues, some universities are opting to reinstate “pen and paper” exams and attempting to boost plagiarism and AI detection software.[4]
Threats and Opportunities
While posing a complex threat to educators, this kind of generative AI offers unparalleled opportunities across a range of industries. Highly specific customisation becomes much more consistently achievable with this technology, as does regular, personal customer engagement and product recommendations. In-depth conversations can be facilitated between consumer and company for the purposes of better explaining products and responding to customer queries.
For marketing and sales purposes, the AI enables highly personalised communication with customers and fast content creation. For those in the legal industry, it can streamline the process of drafting and reviewing legal documents, particularly as this writing form adheres to generally formulaic patterns. It can write code, generate and interpret data and identify anomalies in patterns, greatly impacting departments such as IT.
“It can write code, generate and interpret data and identify anomalies in patterns, greatly impacting departments such as IT.”
Across all industries, this AI tool’s ability to summarise and synthesise masses of information to the standard of a trained professional in a fraction of the time offers huge potential for greater efficiency in organisations. It can analyse customer feedback, create reports and presentations, and answer questions effectively, cutting down the workload for countless workers.[5]
These applications within business reveal a plethora of opportunities opened up by the capabilities of the technology. However, it is not without its weaknesses, nor its dangers.
Formulaic and Unoriginal
One of the current weaknesses of the technology is that its writing style is highly formulaic and unoriginal. It has also been caught fabricating certain facts and references, especially when responding to country-specific questions from regions with which it is unfamiliar. However, it is unlikely that these weaknesses will remain for much longer, given the rapid speed at which the machine learning is able to take place.
From an ethical standpoint, the conversation becomes complicated. It is not difficult to detect the ethical issues, not to mention the educational weaknesses, of the use of the technology within schools and universities. Beyond this, it also raises serious concerns about the biases behind the programming, and questions surrounding fault and blame if it offers information and conclusions that are misinformed or harmful.
It is increasingly becoming the case that the existential questions of yesterday’s science fiction are the tangible concerns of today. The machines we are building are literally smarter than us, having read and processed more content and literature than a lifetime’s worth for a human. Previously, we rested assured that our human qualities set us apart from robots and AI, but with the genuine question of AI consciousness even this seems under threat.[6]
“We are putting a dangerous amount of power into the hands of a machine that is by our standards almost certainly sociopathic”
Princeton University psychology and neuroscience professor, Michael S. A. Graziano highlights how the complex decision-making made by AI does not require consciousness. However, consciousness does account for the human capacities for empathy, social awareness and concern for others. Those lacking in such attributes are referred to as sociopaths, which is precisely what AI would become without consciousness. As suggested by Graziano, we are putting a dangerous amount of power into the hands of a machine that is by our standards almost certainly sociopathic; as a result, we would benefit from AI that is more conscious, not less.[7]
Weaknesses and ethics aside, generative AI like ChatGPT will undoubtedly revolutionise some of those most established processes, frameworks and paradigms across every industry. It offers unparalleled opportunity or immense threat, depending on how each leader and organisation chooses to engage with it.
[1] Graziano, M. S. A. 2023, ‘Without Consciousness, AIs Will Be Sociopaths’, Wall Street Journal, 13 Jan.
[2] Purtill, J 2023, ‘ChatGPT appears to pass medical school exams. Educators are now rethinking assessments’, ABC News, 12 Jan.
[3] Cassidy, C 2023, ‘Lecturer detects bot use in one-fifth of assessments as concerns mount over AI in exams’, The Guardian, 17 Jan.
[4] Purtill, J 2023, ‘ChatGPT appears to pass medical school exams. Educators are now rethinking assessments’, ABC News, 12 Jan.
[5] Chui, M, Roberts, R & Yee, L 2023, ‘Generative AI and other foundation models are changing the AI game, taking assistive technology to a new level, reducing application development time, and bringing powerful capabilities to nontechnical users.’, Quantum Black AI by McKinsey.
[6] Graziano, M. S. A. 2023, ‘Without Consciousness, AIs Will Be Sociopaths’, Wall Street Journal, 13 Jan.
[7] Graziano, M. S. A. 2023, ‘Without Consciousness, AIs Will Be Sociopaths’, Wall Street Journal, 13 Jan.
Article supplied with thanks to Michael McQueen.
About the Author: Michael is a trends forecaster, business strategist and award-winning conference speaker.
Feature image: Photo by Rock’n Roll Monkey on Unsplash