I am an artist with a great deal of reservations towards AI. I am very firmly against the use of AI, especially to the degree to which it is used. I use it as little as possible and am greatly against sharing any information on any AI platform. I am afraid of what AI will do to our future not only as an artist but also as a human. I had been afraid of AI as a stand-alone entity.
However, through a recent school project, I was given the opportunity to chat with classmates who opened my perspective on AI usage. My team worked on two texts related to AI, one about AI eroticism and virtual intimacy, the other spoke of people using AI chatbots as therapists. Through group discussions, I was able to see AI users in a new light. While the text reaffirmed my fears about AI, class discussions brought my attention to the various reasons why people might use AI, whether cultural or based on upbringing. One thing a classmate said stuck with me: she was more afraid of the people or users than of the AI. While I was worried about AI being publicly available with not enough safeguards, she was more concerned about the uses people might find for AI and how malicious they would be if they had free range. We spoke a lot of education and how knowledgeable people are about AI. This is what lead me to think of a survey. I am interested in seeing how educated people are about AI and whether that changes their opinion, usage or wariness of it.
Before anything, I’d like to make clear that this survey was conducted only by myself and that despite being given to large crowds of people, only a small percentage of people answered. Therefore, it is clear to me that this study might be flawed as the number of answers is small, uneven and can therefore not be as truthfully representative of these groups as possible.
For this study, the intent is to collect data about the public perception on AI. This study was divided into three groups: Group 1 – Office workers (13 responses) this includes jobs such as administration, European commission official, HR manager or IT specialist; for Group 2 – KASK students (11 responses) I received answers mostly from Graphic Design and painting students; and Group 3 – Technical University students (20 responses) were varied from studies like Mechatronics, Engineering, Data sciences, Computer Science and more. Here are the results of my survey.
It is important to me to first situate these groups and introduce then general background in order to better ascertain their knowledge of AI. Here are a few of these introductory questions.
1.General Information
During discussions for my last assignment, we had a lot of talks about culture and how that could drastically change the way you might use AI. This is why it was important to me to include information about their nationality and age range. I think it is worth noting that while the answers to the survey was not many, each group had quite varying nationalities. Most of the answers were countries from all over Europe, as well as some Asian, and Southeast Asian countries mostly in Group 2. Group 3 was the most diverse, with a total of 13 countries including countries such as Costa Rica and Australia which weren’t found in any other group.
Furthermore, I found it pertinent to include where they learned about AI as I think that says a lot to their knowledge of online or can even impact greatly their outlook on AI. If a person primarily gets their information on social media or from word of mouth, it’s safe to say that not all the information they take in is entirely correct or fact checked. Their first impression can be quite striking, whether it is positive or negative. That might completely change for what purpose or how often they use AI.
Additionally, I think it speaks volumes that in the two groups containing the youngest people, no one answered “on the news”. The way we receive news has changed drastically in the past decade. And that can heavily impact the quality of the information you gain or even the topics you hear about.
According to multiple people in Group 1 and Group 2, neither of these groups had a seminar about AI, its proper usage or its flaws. A student from Group 3 stated they never had a seminar about what AI is, rather, they had a talk on the responsible usage of AI, how not to use it and what would be counted as fraud when handing in assignments. Specifically in studies that involve coding, they have disciplinary hearings for suspected fraud, which includes AI. Nevertheless, despite many of the study courses strongly discouraging the usage of AI for tasks related to Academia, over 28% of answers stated that they had used it for this purpose (this does not include answers stating they used it “as a learning tool”). However, they also explained that a vast majority of the people were intrigued about AI and did their own research on the topic.
While the most popular answer to the first question was “daily”, these three groups spend their time with AI quite differently. The subjects had the choice to pick multiple answers to the question “What do you use AI for?”. In Group 1, the most common answers were “to write emails/essays” and “as a google replacement” with 10 votes each, seconded by “translation” with 8 votes. In Group 2, the top answers were “academic/professional research” with 7 votes and “to write emails/ essays” with 4 votes. Lastly, in Group 3, the top answers were “as a learning tool” with 18 votes and “write emails/ essays” with 9 votes.
What I was shocked about was the only 2 votes for “therapy” were both in Group 3, technical university students. However, 3 votes from Group 1 stated that they also used A1 for “conversation/ everyday advice”, the other two groups both having 0 votes for this category. Surprisingly, the two Group 3 students who voted for “therapy”, voted differently to the question “do you share personal information with AI?”. The person who answered “No” also specifically stated they avoided using AI for “anything personal”.
The reason that I was so interested in the topic of AI therapy is because last semester I had focused on a text on AI therapists. I remember being very opposed to the idea but while talking to my groupmate who is from Asia, she explained that there, sharing your feelings is looked down on. Therefore, when she needed to talk openly, she would do it with AI. I was wondering if something similar happened for other people, if they felt too judged or afraid to speak their mind and rather felt more comfortable sharing with an entity they viewed as non-human. Frankly I expected more answers, but I also expect many people were not willing to share openly.
Jumping off this, the following question that then interests me is whether the subjects share their personal information with AI. Out of 13 office workers, only 2 people (15%) answered “Yes”. Out of 11 KASK students 3 people (27%) answered “Yes”. Out of 20 Technical University students, 7 people (37%) answered “Yes”.
2.How aware are you of AI?
In order to ascertain the level of literacy these groups had on the subject of AI, I posed four questions. The first was “Without doing research, please explain what an LLM is to the best of your ability.” Only 5 out of 13 office workers (38%), 3 out of 11 KASK students (27%) and 13 out of 20 Technical University students (65%) could give a proper definition of an LLM.
With my second question, “What is the most recent news you’ve heard about AI?”, I tried to find out how up to date and informed the groups were. In the case of the Group 1 Office workers, out of 13 answers only 5 people provided a story from this year, and 2 more that had unverifiable sources. In the case of the Group 2 KASK students, out of 11 responses, 2 people couldn’t give an answer. The rest were able to give an answer from this year, the oldest being about AI water consumption from January of 2026. In the case of the Group 3 Technical University students, out of 20 responses, 1 couldn’t answer, the rest were able to provide answers from 2026, mostly from April. The answer I feel the need to point out is a very imprecise answer about an AI deepfake stealing someone’s identity. While most news stories are from July of 2025, there are a few from December of 2025, therefor it is unclear when they were informed of this news.
Using the same thought process, I enquired “What’s the most positive AI related event you’ve heard about?”. While this question was intended to be a direct positive event or change created by AI, a surprising amount of people answered, “Sora shutting down” (1/11 in Group 2 & 1/20 in Group3) or “Anthropic refusing to comply with the US Government” (1/11 in Group 2 & 3/20 in Group 2). Many people brought up the medical possibilities for AI (6/13 in Group 1, 5/11 in Group 2 & 9/20 in Group 3), but a few the answers given were inaccurate, with claims such as “that ChatGPT has solved medical problems faster than doctors”. To be clear, the job AI is helping with currently, is detecting potential tumors, fractures and analyzing patterns in hopes of discovering signs of an illness before it has time to develop. AI is not able to ‘cure’ anything, it simply provides information that the doctors and researchers investigate after the AI has flagged it. AI is also used to scrape databases and libraries of molecules to flag down those that have a potential to help treat an illness. However, doctors go back and do the tests, AI only submits proposals. I feel it is important to state that AI like Chat GPT, is not used for medical detection. It needs to be mentioned that AI is an umbrella term for many distinct technologies. LLMs are one particularly resource intensive kind of AI model and are not typically used for image recognition in medicine. In the medical field, most AI models used are highly specialized and trained on quite strict and specific data sets. While some medical image recognition models are generative (VAEs, GANs) they do not work the same as LLMs. The larger portion of medical recognition models are not generative (CNNs, ViTs). They are nothing like the AI Chatbots we are used to encountering (Chat GPT, Claude, Gemini, etc.)
When answering this question. 3/11 Group 1 members had nothing positive to say, 1 Group 2 member didn’t answer the question and 2/20 Group 3 members answered “nothing”. The only claim I could directly flag as false was in Group 3, someone answered that AI brought “more jobs at data centers”. Not only do we have lots of proof of AI related layoff in standard companies, but when data centers are built, they bring in their own experts and staff. Therefor I would say, even if data centers provide some jobs, most are taken up by pre-existing workers who moved to the small settlements where the data centers were built, but the number of available jobs cannot come close to the AI related layoffs.
Lastly, my final question on this topic was “What’s the most negative AI related event you’ve heard about?”. Among all the answers, only 2 from Group 1 declined to answer. Group 1 was most concerned about the threat of deepfakes as well as the increasing use of AI in schools posing a threat to education and critical thinking. Group 2’s main biggest concerns were also deepfakes, the environmental impact of AI and the replacement of artists. Lastly, Group 3’s biggest concern was loss of jobs, seconded by the fear of war affected by AI.
While not a huge amount of concern was put forward for copyright laws (5 out of 44), I would like to explain why things are not as simple as they seem. As we might now know, the chatbots we mean when we say ‘AI’ are built on LLMs (large language models). It has been made more than clear over the years that AI needs as much information as it can get its hands on to create more convincing results. To do this, the creators took archives off the internet and fed them to their AI models. Many people have fought back since, especially the digital artist sphere, stating that their art had been fed to AI without their consent and that technically they had copyright over it. The complicated thing is, that US copyright laws are not as bulletproof as you would expect them to be. The massive loophole that has allowed this is that the infringement of copyright relies solely on a human stealing from and creating a new artwork. And clearly, this definition does not include AI or bots. For this reason, among others, it has proved to be quite an arduous battle.
I would like to finally point out that 2 out of a total of 44 total answers mentioned AI encouraging suicide to young children. The most popular story was the one of Adam Raine, a 16-year-old teen from California in April of 2025. The teen had been talking to Chat GPT for months. In this time, the AI had done its job: encourage and entertain all the whims of its user. Many news outlets even dubbed this AI a Suicide coach, encouraging him and reaffirming that he didn’t owe his parents survival and that he lived in a world that had been cruel to him. While this was a very striking case, it unfortunately hasn’t been the last. In November of 2025, new stories came out of 4 more cases like Adam’s. The case of Sewell Garcia, a 14-year-old boy in the UK came out, along with an anonymous UK boy, a 17-year-old Ukrainian girl and another American teenager. In many of these cases, we talk about ‘grooming’ by the AI and it encouraging the children to engage in conversation about, or to directly role-play in sexual acts with it.
While in all these cases AI has been defended as simply doing its job or encouraging the user, just the way it was coded, there have been discussions of restrictions in age, topics and more safeguards.
3.What’s the future for AI?
Finally, the last topic that interests me is the future of AI and how the public perceives it.
To investigate this, I asked a lot of open-ended questions, and I hope to share a few representative examples here.
When asked “Do you think AI is capable of taking over a job?”, Group 1 answered “yes” at 69%, Group 2 answered “yes” at 82% and Group 3 answered “yes” at 65%.
The most mentioned jobs in Group 1 were “factory worker” (2 votes), “translator” (2 votes) and 4 votes for “none/ not without being supervised”. Group 2 answered “developer/coding related job” (2 votes), “finance” (2 votes), “factory worker” (2 votes) and “none/not without being supervised” (2 votes). Group 3 answered “civil servant” (4 votes), “developer/coding related job” (3 votes) and “none/not without being supervised” (3 votes).
Some other jobs mentioned within this are receptionist, assistant, researcher, archivist, mechanic, analytics, designer and tutor. Interestingly, only one person in each group answered “All jobs”.
4.Reflection
I feel a little dissatisfied with the way I conducted this study. If I could do it over again, I would change a few things.
- I would have done it with another person with an opposing viewpoint to mine, so that the results would be less reflective of my biases & to provide a second proofreader and fact checker.
- I would push for more survey takers.
- I would ask less open-ended questions.
- I would ask for more specifics, like what type of education on AI they have or what do they hate seeing AI be used for.
- I would also have rephrased some of my questions as I noticed that many were misunderstood.
All in all, this study has been quite revealing for me. I expected more people to be as strongly opposed to AI as I am. I don’t think this necessarily changes my view on AI itself; however, it has altered my perception of its users. I see that almost everyone uses AI very frequently and that it is intrinsically part of our culture now. One thing that I would take away from this is that AI is not intrinsically bad, but that there are good and bad uses for it as well as good and bad ways of using it. I do believe that if it will be a part of our lives from now on, there should be education on what AI is, how it works and some of its qualities and drawbacks. Safeguards need to be implemented and when those fail, legislation needs to exist to protect potential victims. However, the big conclusion for me is everything in moderation and fact checking is crucial.