AI Can’t Replace Us: What Youth Really Think About AI
15 young people across high school and college share their views on AI’s role in society
Researched and Written by Bomi Akarakiri, YPA Research and Policy Director.
Toplines
Respondents acknowledged AI's role in society as a helper tool to improve efficiency, but were more skeptical about the longer-term impacts it may have.
Respondents were more critical of companion AI in comparison to functional. They agreed that AI cannot substitute human relationships and should have limited human-like features.
Respondents agreed that greater restrictions should be placed on minors around companion AI use. 12/15 were in favor of a complete ban whilst the remaining proposed content restrictions.
Respondents thought that governments, schools and developers should be held more accountable than individuals for AI issues and its potential harm.
Introduction
Between Monday July 7, 2025 and Friday July 18, 2025, YPA posed questions for high-school and college students to report their views on Artificial Intelligence (AI). The survey was conducted remotely via Google-meet as well as in a Google form. There were 15 responses (8 participating in the virtual meeting and 7 for the Google form). Though all were YPA members, which indicated an interest in youth advocacy, the results (shown below) revealed that only 3/15 of the participants had been involved in AI-related initiatives on civic engagement and economic opportunity. Therefore, YPA’s survey represents young people who are not involved in tech policy work.
Demographics
Total number of respondents: 15
Race: AAPI 6/15, White 6/15, Hispanic 2/15, Black 1/15
Gender: Male 7/15, Female 8/15
No. of respondents in high school: 10/15
No. of respondents in college: 5/15
Age range of respondents: 14 - 22
Functional vs Companion AI
To gain a more distinguished understanding of views towards AI, YPA categorized AI into two categories. Functional AI was defined as systems designed to perform specific tasks whereas companion AI are systems designed to provide emotional, social, or interactive relationship-like connections to humans. With slight variation depending on whether respondents attended the virtual meeting or submitted using the form, the questions covered AI use in everyday lives (which included any advocacy involvement and awareness of use by friends), opinions towards AI, who should be held accountable for misuse, thoughts on policy reform (restrictions, bans, disclosures) and what respondents envisioned healthy AI to be.
Use and experience
Consistent with their stage of life, 13/15 respondents explicitly stated that they use AI for academic purposes with direct mention of the words “school” “education” “class” and “essays”. Similarly, 9/15 respondents said they have friends who use AI- citing similar purposes. One student in particular shared that they watched friends become addicted to character AI. But this was an isolated view as others explicitly said they were not concerned about their friends’ use. When asked if respondents had any AI advocacy experience, the majority of respondents, 10/14, admitted that they have not (one student did not answer this question). YPA was responsible for three respondents’ involvement; as field directors, they had the opportunity to lobby for social media legislation. Also, one student had attended the Stanford AI4ALL program.
Opinions and Concerns
Nuances began to develop when questions transitioned from AI use to policies around functional and companion AI. Overall, sentiment was more negative for companion AI than functional, which was supported by the 15/15 consensus that AI cannot substitute human relationships, and the 14/15 agreement that human-like features of AI should be limited. When asked “Do you have a differing opinion of functional AI vs companion AI”, one respondent said that functional AI is more widespread but has less severe consequences, however another respondent conveyed greater worry for functional AI BECAUSE it is more widespread. There were also varying beliefs on whether or not AI can substitute for professional help. Some (7/15) thought it depended on the profession and depth of assistance provided, others (2/15) predicted that it could in the future and the rest asserted no; AI cannot substitute professional help. Additional concerns that respondents had regarding AI were to do with job security, the environment, increased dependency and the lack of regulation.
Regulation
The question “Who should the primary responsibility fall on to address problems with AI” was asked to the 8 respondents who joined via Google meet. Though respondents had different views on whether developers, schools or the government should primarily be held accountable for any potential harm experienced by AI, all showed more compassion towards individuals who were consistently ranked the least responsible. Reasons for this included the lack of AI education and how vulnerable AI companion users are perceived to be. YPA then asked if respondents supported a ban on companion AI use for minors. Only three respondents opted for content regulation in lieu of a complete ban, everyone else voted yes.
Conclusion
The survey concluded with the question “What do you believe healthy AI use looks like?”. Respondents endorsed AI’s existence in society as a secondary tool for tasks such as editing, research and everyday life. The parameters for this was that as long as AI does not increase its human-like notions or lead to dependency, interactions can remain healthy.
Quotes from Students
“Calibre of academics are going to be changed by reliance on AI”
“Losing critical thinking skills but definitely has its place on making things more efficient”
“Companion AI tells you what you want to hear which is a scary form of reassurance”
“AI is in the process of revolutionizing our society which is dangerous without regulation”
“AI in class is detrimental for our intelligence as a generation.”
“Using AI for everything becomes a habit. You think less.”
“Sad to see people disregard their own capacity in favor of AI.”
“Nervous for future work integrity.”
“AI (is) harmful at best.”
“Saddening to see people are relying on AI to recreate human brilliance”
“Relying on AI for human connection and mental help is a very slippery slope and can be very dangerous for youth”
“The increase in AI use creates user dependent outcomes.”
 | A guest post by
|