.png){kind=logo}
Executive Summary
Research findings indicate that Kogod students are actively using AI, employers are increasingly asking about AI capabilities, and students want more structured preparation to use AI effectively, ethically, and competitively in the workplace.
Kogod student survey data shows that AI skills are rapidly becoming a workplace expectation rather than an optional advantage.
The share of undergraduate and graduate students who reported being asked by potential employers about their ability to use AI increased sharply over the past three years, rising from 11.6% in 2024 to 32% in 2025 and 42.6% in 2026. This trend suggests that employers are increasingly treating AI fluency as a practical job-readiness skill.
Student AI adoption is also widespread.
More than 80% of students reported using AI tools for academic work in the past six months, with ChatGPT and Perplexity* emerging as the two most used platforms. Gemini ranked as the third most used platform, but remained significantly behind at approximately 37%. When asked to identify their favorite AI model, undergraduates selected ChatGPT, while graduate students showed a stronger preference for Claude.
*At the time of this study, all students surveyed had access to Perplexity AI Pro through Kogod’s enterprise partnership with Perplexity.
Across all three years of survey data, brainstorming remained the number one use case for AI.
This indicates that students most commonly rely on these tools to generate ideas, organize thoughts, and begin assignments or projects. When choosing which AI model to use, students prioritized accuracy above other factors.
At the same time, students continue to express concern about responsible AI use.
Nearly half of students indicated some level of concern about academic integrity and the use of AI. These findings highlight the need for clearer guidance, stronger academic norms, transparent expectations around appropriate AI use, and teaching students explicit judgment and critical thinking skills to evaluate AI outputs.
The data also reveal that students want more advanced and practical AI training.
Prompt engineering remains the top AI skill students want to learn, but coding has now become the second most desired skill, suggesting a shift from basic AI usage toward more technical and career-oriented applications. Other high-interest areas include model strategy, workflow optimization, research, and general AI knowledge/tools.
2026 Authors
| Creation of Survey | Christina Eid, Caron Martinez, Henry Ho, Gwanhoo Lee |
| Collection of Data | Christina Eid, Kogod Faculty and Staff |
| Analysis of Results | Christina Eid, Henry Ho |
| AI Report | Christina Eid, Caron Martinez |
Methods
The surveys were conducted from February 15, 2024, to March 26, 2024; from April 10, 2025, to May 7, 2025; and from March 29, 2026, to April 7, 2026. A Google Form was used in 2024 before moving to a QualtricsXM survey in 2025 and 2026. Professors from a range of Kogod core and elective courses posted a link to the survey so that students could complete it in class. The respondents were fairly equally distributed from first years to graduate students and represented all Kogod majors. A self-report, anonymous survey was used to collect the data. It was determined that in order to receive the most honest and accurate results possible, the survey would be anonymous. 609 responses were collected over the three years, but a sample size of 483 over the three years was used for most of the data in this report (when non-Kogod majors and undeclared majors were filtered out).
All data is self-reported. Furthermore, a key limitation of this study is its sample size and its focus on students from only American University’s Kogod School of Business, which may limit the generalizability of the findings to broader student populations.
Throughout the survey, students were informed that the survey results are anonymous, allowing students to answer freely and honestly. Qualitative data analysis was performed on the open-ended questions. All open-ended questions were categorized using a custom app made with Google Gemini. To ensure accuracy, responses were analyzed by a person without the aid of AI, allowing the comparison of the results from both methods.
KEY TAKEAWAY: WEEKLY AI USAGE HAS SKYROCKETED
We asked: “How many times a week do you currently use AI for school, internship, or employment-related tasks?”
| Weekly Usage | 2024 | 2025 | 2026 | Δ 2024–2026 |
|---|---|---|---|---|
| 0 times | 17.1% | 6.4% | 4.3% | ↓74.9% |
| 1–3 times | 46.6% | 29.6% | 26.6% | ↓42.9% |
| 4–7 times | 23.3% | 27.2% | 30.0% | ↑28.8% |
| 8–10 times | 6.8% | 12.0% | 10.1% | ↑48.5% |
| 11+ times | 6.2% | 24.8% | 29.0% | ↑367.7% |
KEY TAKEAWAY: TOP FIVE STUDENT USES OF AI TOOLS REMAIN RELATIVELY CONSISTENT, WITH SOME NEW ADDITIONS
We asked: “How do you use AI?”
| Use Case | 2024 | 2025 | 2026 |
|---|---|---|---|
| Brainstorming | 67.8% | 75.2% | 75.8% |
| Studying for exams | 1.4% | 59.2% | 66.2% |
| Summarizing | 50.0% | 65.6% | 62.3% |
| Explaining concepts | 48.6% | 58.4% | 58.5% |
| Finding sources | 17.1% | 48.0% | 54.1% |
| Writing cover letters, resumes, CVs | 29.5% | 40.0% | 38.6% |
| Analysis | 22.6% | 40.8% | 39.6% |
“This question asks you to select an answer based on how much you agree with the following statements.”
| Statements for Level of Agreement (1=Disagree, 5=Agree)* | 2024 Avg. | 2025 Avg. | 2026 Avg. |
|---|---|---|---|
| I feel like I have the skills/knowledge to use AI for academic purposes. | 3.72 | 4.49 | 4.46 |
| I feel prepared to use AI for job/internship related purposes. | 3.36 | 4.34 | 4.26 |
| My instructors have adequately discussed their AI policy specific to their course(s). | 3.59 | 4.23 | 4.28 |
| AI has improved my understanding of course material. | 3.75 | 4.41 | 4.19 |
| AI allows me to complete assignments more efficiently. | 3.80 | 4.37 | 4.35 |
| I have concerns regarding the use of AI and academic integrity within Kogod. | N/A | N/A | 3.33 |
*1=Disagree, 2=Somewhat Disagree, 3=Neither Agree Nor Disagree, 4=Somewhat Agree, 5=Agree
If respondent answered "yes" to previous question, we asked: “What specific skills/concepts related to AI have potential employers asked you about?”
Top categories based on 2026 qualitative responses:
- General AI & data analysis questions: Asked over 50% of the time
- Practical AI use: Prompt engineering, research, data analysis, workflow efficiency, content generation, coding, development, and job-specific applications
- Responsible use: Where students draw the line between using AI as a tool and over-relying on it
We asked: “Which AI language models have you used for academic purposes in the last 6 months? Select all that apply.”*
*At the time of this study, all students surveyed had access to Perplexity AI Pro through Kogod’s enterprise partnership with Perplexity.
| Model | 1st Year | 2nd Year | 3rd Year | 4th Year | Grad | Overall |
|---|---|---|---|---|---|---|
| Perplexity AI | 87.9% | 96.4% | 78.9% | 86.1% | 82.1% | 86.0% |
| ChatGPT | 68.2% | 78.6% | 84.2% | 94.4% | 76.9% | 78.7% |
| Gemini | 34.8% | 39.3% | 31.6% | 36.1% | 46.2% | 37.2% |
| Claude | 16.7% | 25.0% | 34.2% | 30.6% | 71.8% | 33.8% |
| NotebookLM | 15.2% | 46.4% | 18.4% | 19.4% | 41.0% | 25.6% |
| Copilot | 19.7% | 21.4% | 18.4% | 16.7% | 20.5% | 19.3% |
| DeepSeek | 3.0% | 7.1% | 13.2% | 5.6% | 12.8% | 7.7% |
| MetaAI | 1.5% | 3.6% | 2.6% | 2.8% | 0.0% | 5.3% |
| Grok | 3.0% | 3.6% | 5.3% | 5.6% | 10.3% | 1.9% |
| Other | 1.5% | 0.0% | 0.0% | 0.0% | 0.0% | 2.6% |
We asked: “Of the AI language models you have used for academic purposes in the last 6 months, please select your favorite.”
| Model | 1st Year | 2nd Year | 3rd Year | 4th Year | Grad | Overall |
|---|---|---|---|---|---|---|
| ChatGPT | 34.8% | 42.9% | 34.2% | 44.4% | 25.6% | 35.7% |
| Perplexity AI | 40.9% | 28.6% | 26.3% | 22.2% | 23.1% | 30.0% |
| Claude | 16.7% | 10.7% | 21.1% | 8.3% | 30.8% | 17.9% |
| Gemini | 0.0% | 7.1% | 10.5% | 8.3% | 10.3% | 7.2% |
| NotebookLM | 1.5% | 10.7% | 5.3% | 11.1% | 7.7% | 6.3% |
| Copilot | 1.5% | 0.0% | 0.0% | 0.0% | 0.0% | 0.5% |
| DeepSeek | 0.0% | 0.0% | 0.0% | 2.8% | 0.0% | 0.5% |
| Grok | 0.0% | 0.0% | 0.0% | 0.0% | 2.6% | 0.5% |
| Other | 1.5% | 0.0% | 0.0% | 0.0% | 0.0% | 0.5% |
We asked: “What is the single most important factor influencing your choice of an AI tool?”
| Factor | 1st Year | 2nd Year | 3rd Year | 4th Year | Grad | Overall |
|---|---|---|---|---|---|---|
| Accuracy (less bias/hallucinations) | 40.9% | 35.7% | 44.7% | 31.4% | 38.5% | 38.8% |
| User experience (ease of use) | 13.6% | 7.1% | 15.8% | 20.0% | 17.9% | 15.0% |
| Specific features (e.g., multimodal inputs, data analysis, coding) | 4.5% | 17.9% | 10.5% | 22.9% | 7.7% | 11.2% |
| Speed | 10.6% | 14.3% | 5.3% | 5.7% | 7.7% | 8.7% |
| Cost | 7.6% | 7.1% | 5.3% | 2.9% | 15.4% | 7.8% |
| Familiarity with the AI tool | 9.1% | 7.1% | 10.5% | 11.4% | 0.0% | 7.8% |
| Other | 9.1% | 3.6% | 7.9% | 2.9% | 5.1% | 6.3% |
| Privacy & data | 4.5% | 7.1% | 0.0% | 2.9% | 7.7% | 4.4% |
We asked: “Are there specific skills/concepts related to AI that you would like to be taught in class?”*
*Responses to this question and data shown in the chart below it come from the full dataset of 609 respondents rather than the smaller sample that filtered out non-Kogod majors and undeclared majors.
Top reasons based on 2026 qualitative responses:
- Prompt Engineering & Model Strategy
- The most requested skill
- How to craft better prompts, choose between different models (like Claude vs. Perplexity), and understand the logic behind AI responses
- Coding & Technical Development
- Strong interest in “Vibecoding,” building AI agents, webscraping, and using AI to assist in software/app development
- General AI Knowledge, Ethics, & Tools
- Understanding the fundamentals of how AI works, its environmental impact, and staying updated on the latest AI tools and trends
- Productivity & Workflow Optimization
- Focuses on making work and assignments more efficient, automating repetitive tasks, and integrating AI into daily workflows (e.g., using AI with Excel)
- Research & Academic Learning
- Using AI for academic research, summarizing complex concepts, and assisting with essay writing and studying
- Business & Professional Applications
- Practical applications for finance, data analytics, resume screening, and building AI-driven businesses
- Creative & Content Creation
- A smaller but notable interest in generating images, videos, and design work
| Skill Area | 2025 | 2026 |
|---|---|---|
| Prompt Engineering & Model Strategy | 37.5% | 33.0% |
| Coding & Technical Development | 3.6% | 19.8% |
| General AI Knowledge, Ethics, & Tools | 16.1% | 17.0% |
| Productivity & Workflow Optimization | 10.7% | 12.3% |
| Research & Academic Learning | 19.6% | 9.4% |
| Business & Professional Applications | 8.9% | 5.7% |
| Creative & Content Creation | 1.8% | 0.9% |
| Other | 1.8% | 1.9% |
We asked students to provide qualitative responses about any concerns they have regarding AI.
| Concern | 2024 | 2025 | 2026 |
|---|---|---|---|
| Cognitive & Educational Devaluation | 25.2% | 30.8% | 30.2% |
| Academic Integrity & Loss of Originality | 31.1% | 10.8% | 10.1% |
| Adverse Career Impact on Future of Work | 13.4% | 13.8% | 20.2% |
| Technical Accuracy & Unreliability | 9.2% | 20.0% | 18.6% |
| Ethics, Privacy & Governance | 14.3% | 13.8% | 13.2% |
| Social & Environmental Responsibility | 5.9% | 9.2% | 6.2% |
| Other/General | 0.8% | 1.5% | 1.6% |
We asked: “Have you ever used AI as a shortcut or substitute for your own work, rather than as a supplemental tool?”
| Response | 2025 | 2026 |
|---|---|---|
| “Yes” | 46.5% | 43.5% |
| “No” | 53.5% | 56.5% |
If respondent answered "yes" to previous question, we asked: “In which of the following situations have you felt inclined to use AI as a shortcut or substitute for your own work? (Select all that apply.)"
| Situation | 2025 | 2026 |
|---|---|---|
| Time constraints | 85.4% | 79.5% |
| Pressure to achieve good grades | 62.5% | 53.4% |
| Assignment/material seemed like “busywork” | 64.6% | 56.8% |
| Assignment/material had a high level of difficulty | 54.2% | 51.1% |