The “advanced search engine” for some and an “overestimated tool” for others. Is Generative AI developing too fast for humanity to keep up with?
ChatGPT has quickly gained popularity and become an easily accessible tool for University students. However, there is a wide spectrum of opinions on whether it’s a simply a helpful tool or secretly harming humanity. Antara Verma explores how opinions and predictions about ChatGPT and other forms of generative AI differ among four students with different views, one of which works in chatbot development
ChatGPT has rapidly grown to be one of the most used AI chatbots since its release by OpenAI in late November of 2022. It did not take long to attract users and in only 2 months, over 100 million people had created an account. Its accessibility and efficiency keep attracting more and more users and there are currently over 180 million users worldwide. This is unsurprising since anyone can create an account for free and then type in a question, or prompt and it can produce almost anything from writing essays to coding computer programs. This tool has become known to many university students and has facilitated their researching, studying, and even writing processes. Naima (22), who uses ChatGPT on a every week, uses it for “Old English Translations, to summarise articles, and to get some ideas started for essays”. While generating ideas and translating are quite common uses among students, some even use ChatGPT as “an advanced search engine”, as Harry (22), who uses ChatGPT daily, calls it.
Despite these benefits, there are a lot of ethical concerns that come with the use of generative AI and at university one of the main concerns is plagiarism. Before the widespread of ChatGPT, professors “only” had to worry about students copying their work from pre-existing papers and research. This had been quite easy to recognise with simple plagiarism tools and also knowing what student’s writing looks like. However, generative AI has added a new dimension, making it easier than ever for students to plagiarise. All they have to do is type in “Write an essay with the title X” or “Solve this physics problem for me” and that’s all it takes to complete an assignment. There is no doubt that most professors are not happy with the spread of generative AI, since it makes them correct any work pointless. Parents are similarly not content with paying for their child’s tuition only for them to not learn anything at all. Angelo (21), who uses ChatGPT 3-5 times monthly, expresses his biggest concern, which is that “People get too lazy to do their own research or analyses”. Laziness is a big concern that arises with generative AI. If students can complete any assignment in a matter of seconds, how will they ever learn anything for their degree? Naima expressed her concerns by saying that she feels guilty when using ChatGPT and explained: “I know I should be able to do the work myself and it is often a shortcut. Also, getting used to summarising texts makes me feel guilty and my attention span to read a paper of 30 pages that could potentially be really interesting is not there anymore because of the option to use ChatGPT”. Her answer focuses on the issue of becoming too dependent on generative AI and she elaborates on why this could be a problem: “Our generation, including me, is getting too used to using ChatGPT and this might have negative effects and consequences in the future. For instance, we are already deviating from the norm of reading texts entirely and finding the shortcuts [by asking ChatGPT to prove a summary]. So, how will we perform in jobs when we have to constantly look things up on AI chatbots?”
This question is precisely why universities around the globe have had to add many regulations surrounding ChatGPT. Most of them have created a list of guidelines on their university’s homepage that include a so-called “Declaration of Authenticity”. This is a document that must be signed to declare that one has done the work themselves before submitting assignments. Apart from what was declared, professors now use programs like Turnitin to detect any use of AI. If Turnitin or any other programs detect the use of AI, there are severe consequences for the student involved. While there are many issues concerning plagiarism and ethics, many students prefer to see the benefits that ChatGPT has for their studies. For instance, when asked about having any concerns with the use of ChatGPT at university, Harry answered: “No, because university is all about learning and as long as you learn and understand what you’re writing and don’t just mindlessly use it, you still learn something”. He further explained how it helps him to “create a general structure [for essays] and put him in the right direction”.
Despite admitting to using ChatGPT, the interviewees agreed that there are problems while using AI. While Naima and Angelo similarly expressed their concerns about Gen Z’s ability to navigate the world after university by doing their own work, Harry was more focused on ethical concerns that arise when ChatGPT is misused by certain people to justify unethical things. He explained: “I read that in the US, some racist people who say that Black people commit a disproportionate amount of crime, ask ChatGPT about it by changing the question slightly, like ‘if there were 100 sheep and 13 were committing crime, AI would say you should remove these sheep’”. Hence, the way that ChatGPT is crucial for its outcome. People with intentions to harm others can trick ChatGPT into saying discriminatory things that are harmful to certain groups of people. This shows how ChatGPT has already come a long way from when it was first launched, for example, in its beginning stages one could only type in a prompt or question and not upload any files for it to read. Currently, users can upload a wide range of files, like PDFs, JPEG, PNG etc. and the chatbot can read, analyse, and/or critique the content document within seconds. A more recent update allows users to search the web using ChatGPT. This essentially renders search engines valueless, since ChatGPT now provides links to articles and websites of any kind that are very specific to what one needs. This explains why many interviewees found that it was useful for research, like Harry who said that he uses it to “help with research for papers” and Naima to help with her Old English translations.
This advancement that ChatGPT has gone through indicates that it will only continue to grow and develop new features. Harry believes that “ChatGPT will be even better at writing academic texts, which I think will be a problem because I think AI will get to such a point where it is indistinguishable from human writing and become undetectable. So, there would be some academic problems”. Similarly, Naima responds: “The new version of ChatGPT will be much smarter and bigger than before, and the technology is going to improve and get better and better. The answers are going to be more human-like and more correct.” In contrast, Angelo responded: “I think we are overrating ChatGPT at the moment, it is not omnipotent”. He thinks that “a limit will be reached with tools like ChatGPT, and people will get more and more used to it and develop a better knowledge of spotting the writing style of AI in general. There will also be more and more guidelines concerning AI at Uni or in the workplace”. Hence, it is clear that with time generative AI will advance, but it also means that people will have more time to be exposed to it and start recognising its patterns. This mix of answers reflects the diversity of opinions about ChatGPT.
For this article, Serafin (23), who is a full-time student, is also responsible for the AI tool development of an international company, shared some insights on generative AI that he has gathered from experience. When asked about how generative AI will advance in the future, he says: “There are speculations that we will hit a point where ChatGPT and generative AI generally don’t get better”. Therefore, what Serafin says is in touch with the end of the spectrum where people do not think that generative AI has the ability to become extremely powerful. However, he continues to explain that “The next big step is the agentive AI that is the next thing that people want and are developing. It gives your AI autonomy on what it wants to do. It is no longer just generating texts; it does tasks for you. Recently, there was an AI launch from Devin that can build and compile using codes without you having to do anything. There has been some headway done there but there is talk that we will soon have reached a ceiling of how far AI can go”. This form of agentive AI is likely to make it even easier for University students to let their chatbots do their work for them, which was a concern expressed by students above.
One of Serafin’s biggest concerns with generative AI is ethical, he explains: “While it is a great tool, but it gives you so many abuse angles. It has been killing the internet with much more misinformation out there. Companies like google use AI in their searches so there is a lot of information on the internet that you don’t even know that it is AI”. Angelo also mentioned this issue by saying “There’s also the ethical concern; we don’t know how trustworthy the information really is that ChatGPT provides us with”. The range of opinions surrounding generative AI, highlight its relevance and controversy at the moment and that its future is still unknown. The last question that everyone was asked was whether they think humanity is better off with generative AI to see if students think this huge leap in technology was for the greater good or not, since it will ultimately be relevant for their futures.
There were some clashing beliefs for this last question and Harry responded: “I think we are better off but at the end of the day it doesn’t matter because it will stay here forever and every technological advancement, although there are drawbacks, it is there to help in some way”. While Harry is very pro-ChatGPT, Naima took more of a neutral stance and said: “It has positive and negative sides”. On the one hand Naima appreciates the technological advancement and says: “It is a very fascinating thing to see what humans have manages to create and it can be very helpful if you don’t misuse it. It can give one idea and deliver a summary. And if you can’t find an answer online anywhere else, you can use this tool”. On the other hand, she notices the flip side and mentions: “It feels a bit scary how many people just totally use AI to do their work, like to write an entire essay. You can cheat more easily now and do less and less work. In that way we would also be better off without it”. Similarly, Serafin appreciates all that generative AI can accomplish; however, he thinks that “It is developing too fast for humanity to figure out to use and regulate it correctly, so that more good than harm are being done. We don’t have that much control over what it says, and people don’t know how to regulate it. It has developed too quickly for it to be fully safe, which is the case for many technological developments”. At this end of the spectrum, Angelo says: “It is healthier for human beings to do tasks themselves that stimulate their brain and although it might be tedious at some point, your brain will thank you in the long run. It is better if we do not let AI tools like ChatGPT take over our lives”.
