StudyAgent Research Finds: 40% of Students Have Test Anxiety Problems
StudyAgent Research Finds: 40% of Students Have Test Anxiety Problems
This article presents StudyAgent’s survey on student AI use, covering exam prep, assignments, top tools, spending habits, and the growing role of AI in daily academic life.

Nov 28, 2025

Research
6 min read

Table of contents
Key Takeaways:
- 40% struggle with exams, 39% with math/programming.
- 60% use AI for essays, 56% for research papers, 46% for coursework.
- 58% grammar checks, 42% essays, 39% summaries, 32% outlines, 30% math.
- Top tools: ChatGPT (64%), Grammarly (39%), Kortext (36%), Google Translate (35%), Copilot (15%).
- Biggest growth: Text generation up from 30% → 64%, coding 6% → 15%.
Artificial Intelligence is now woven into everyday study habits, from quick fact-checks to full essay drafts. To see how this shift is reshaping academic work, StudyAgent ran a U.S. survey on October 6, 2025 through Prolific, reaching 2,083 college students who actively use AI chatbots.
Motivations and concerns stand out clearly. The top reasons for using AI include saving time (51 %), improving work quality (50 %), and getting instant support (40 %). At the same time, worries persist: 53 % fear being accused of academic misconduct, 51 % cite false or “hallucinated” results, and 31 % say their institution discourages or bans AI.
This article reports what we found and why it matters. After outlining the survey design, it examines key insights on how students use AI, their main challenges, and the motivations and concerns that guide their choices.
Methodology
The survey combined internal data analysis, desk research, and a structured questionnaire.
- Platform & sample: Prolific, 2,083 U.S.-based students using AI chatbots.
- Devices: mobile, tablet, and desktop.
- Timing: October 6, 2025.
Questions focused on study habits, frequency and purpose of AI use, preferred tools, spending patterns, and reasons for or against AI adoption. Results were compared with third-party studies to verify trends and provide broader context.
Main Research Insights
The findings show how deeply AI is embedded in learning. 86% of students already rely on it, and more than half use AI daily. The next sections explain the key challenges students face, the tasks they automate, and the tools they depend on.
40% of Students Struggle with Exam Anxiety and Prep

Exam preparation and test anxiety are the biggest struggles for students. About 40% said timed exams are their hardest challenge, while 39% pointed to math, statistics, and programming tasks, which are especially common in STEM fields.
Other major difficulties include finding reliable sources for assignments (29%), combining information from multiple readings (27%), and organizing essays or projects (27%). Around one in four students also struggle with understanding lectures (26%) and creating proper citations (25%).
Collaboration and writing add more stress. About 22% said group projects or paraphrasing text are difficult, and 21% have problems with formatting styles such as APA or MLA. One in five students pointed to unclear or low-quality teaching as an issue (20%), while others struggle with presenting data clearly (18%) or following grammar and academic style rules (15%).
Together, these results show how wide-ranging student challenges are, from test-related stress to writing and research skills, all of which influence academic performance.
Homework Types Students Use AI For

AI has become a mainstay in academic work, especially for writing-heavy tasks. The majority of students use it for essays (60%), followed closely by research papers (56%) and coursework (46%). Literature reviews also see strong reliance at 30%, while about a quarter of students turn to AI for discussion posts (28%), reports (27%), reflective writing (26%), PowerPoint presentations (26%), and case studies (24%).
More specialized uses, such as theses (18%), SWOT analyses (12%), capstone projects (12%), and dissertations (10%), show that AI’s role extends well into advanced study levels. Even in smaller numbers, these figures highlight how AI is becoming embedded across the full range of assignments, from entry-level coursework to professional research.
Academic Uses of AI Tools

AI is now a core part of learning, supporting students at nearly every stage of their academic work.
External studies show that the most common uses include understanding concepts (56%), researching assignments and projects (52%), generating first drafts (46%), writing essays and assignments (41%), and preparing presentations (36%). Students also report relying on AI for exam prep (36%) and homework checks (33%).
Our own StudyAgent survey reveals a very similar pattern. The top use is grammar checking (58%), followed by writing essays and papers (42%), summarizing long texts (39%), and generating ideas or outlines (32%). Other significant tasks include solving math problems (30%), style editing (26%), and fact-checking data (26%).
Taken together, both external studies and our survey show that students most often turn to AI for:
- Grammar checking
- Explaining concepts
- Writing essays and papers
- Summarizing text
- Generating ideas and outlines
- Editing style
These findings highlight that AI has become a steady partner in writing, problem-solving, and managing complex coursework.
AI Tools Students Rely on for Studies

The survey shows a sharp rise in student use of AI between 2024 and 2025. The most common purpose is generating text with ChatGPT (64%), more than double last year’s figure of 30%.
Students also turn to Grammarly (39%) for enhancing and editing writing, and to Kortext (36%) for summarizing, note-taking, or quizzing textbooks. Google Translate (35%) is widely used for language support, reflecting the need for cross-language learning.
Other important uses include speech-to-text transcription with tools like YouTube captions (24%), and media generation with DALL·E (19%) for images, videos, or audio. For technical work, students increasingly rely on Julius AI for data analysis (15%) and GitHub Copilot for coding (15%), both showing notable growth compared to 2024.
Specialized software also appears in the mix, with 11% using AI for discipline-specific tasks such as medical tools.
The shift is clear: in 2024, over a third of students reported not using AI at all. By 2025, the majority are actively integrating these tools into their daily studying, writing, and technical projects.
Final Takeaways
The survey makes one thing clear: AI has moved from being optional to becoming a core study tool. Students use it to manage stress, save time, and concentrate on deeper learning and creativity. It is no longer just a shortcut but an essential part of modern education.
Education leaders are noticing the shift. As UNESCO observed in 2024, “Artificial intelligence has the potential to fundamentally transform education systems, but it must be steered with human values at its core.” Likewise, the OECD recently stated, “AI can support more personalized learning and free teachers to focus on higher-value tasks, provided its use is transparent and responsible.”
Overall, students are not treating AI as a passing trend. They are folding it into their study habits to make it a steady part of modern education.
“AI has changed how students study, but it hasn’t replaced the effort behind learning. Real progress happens when students use it to think better, not less.” - Kateryna Bykova, VP of Content Marketing at StudyAgent.
Sources:
- This research is based on an internal survey of the StudyAgent team, which collaborated with 2,083 students.
- Digital Education Council. (2024). Digital Education Council Global AI Student Survey 2024 (Report). https://26556596.fs1.hubspotusercontent-eu1.net/hubfs/26556596/Digital%20Education%20Council%20Global%20AI%20Student%20Survey%202024.pdf?utm_medium=email&_hsenc=p2ANqtz--b6KpvKWDBx-HpwG4komPEwKJR8a_TCBtzo4keq1HBfE5ig9KNzWI6kR88e_H0o1G2Prkz3rM9Hb7iNSYQsAOJaKbV2CPtSOLKeFxsjQgD6ylnm-8&_hsmi=92199303&utm_content=92199303&utm_source=hs_automation
- Alshareef, N., Giga, S., & Fletcher, I. (2025). Test anxiety, emotional regulation and academic performance among medical students: a qualitative study. Medical Education Online, 30(1). https://www.tandfonline.com/doi/full/10.1080/10872981.2025.2505177
- Microsoft. (2025). A Microsoft Special Report 2025 AI in Education. https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/bade/documents/products-and-services/en-us/education/2025-Microsoft-AI-in-Education-Report.pdf
- Fioriti, C. M., Pizzie, R. G., Evans, T. M., Green, A. E., & Lyons, I. M. (2025). Math anxiety and arithmetic learning: Evidence for impaired procedural learning and enhanced retrieval learning. Journal of Experimental Psychology Learning Memory and Cognition. https://psycnet.apa.org/doiLanding?doi=10.1037%2Fxlm0001453
- Črček, N., & Patekar, J. (2023). Writing with AI: University Students’ Use of ChatGPT. Journal of Language and Education, 9(4), 128–138. https://jle.hse.ru/article/view/17379/16438


