Survey of Computer Science Faculty: Use of Artificial Intelligence

This study that explores how computer science faculty across the United States are adopting and utilizing AI tools in their research, teaching, and administrative work. Drawing on detailed survey responses, the study provides an in-depth look at the prevalence, intensity, and future plans for AI tool usage among faculty members. The report covers a wide range of AI applications, from code assistants to general-purpose large language models (LLMs), and analyzes adoption patterns by institution type, faculty rank, scholarly focus, and demographic factors. Key Findings: ChatGPT Dominates Usage: Nearly two-thirds (66.67%) of surveyed faculty have used ChatGPT, making it the most widely adopted AI tool in the sample. Usage is high across all institution types and ranks, with particularly intensive use among data science, AI, and younger faculty. The average time spent with ChatGPT in the past month was 273.76 minutes, with some reporting up to 3,000 minutes. Gemini and Claude Form a Second Tier: Gemini (30.43% usage) and Claude (26.09%) are the next most popular general-purpose LLMs, often used alongside ChatGPT. Gemini is especially prevalent in community colleges and among interdisciplinary scholars, while Claude is favored in top-ranked programs and by AI/ML faculty. GitHub Copilot Leads Code Assistants: Among code-focused tools, GitHub Copilot stands out with 17.39% adoption, far surpassing other code assistants like Code LLaMA (8.70%) and niche tools such as StarCoder, Replit Code LLM, and Codeium, each used by less than 6% of faculty. Intensity of Use Varies Widely: While ChatGPT users report high engagement, most other tools show low average time spent and zero median usage, indicating that regular use is concentrated among a small subset of faculty. Multi-Tool Workflows and Future Plans: Faculty anticipate continued multi-tool use, with ChatGPT and Claude favored for reasoning and summarization, Copilot and Claude Code for programming, and Gemini for multimodal tasks. Some respondents express caution or ethical reservations about AI adoption. This essential resource offers actionable insights for academic leaders, technology vendors, and policymakers seeking to understand and support the evolving role of AI in computer science education and research.