USC Embraces AI on Campus with OpenAI Partnership, Balancing Innovation and Integrity

USC backs AI in the classroom-while addressing real risks

Artificial intelligence is moving into coursework, grading support, and student study habits. At the University of South Carolina, faculty and staff are embracing AI tools to improve efficiency and learning outcomes, while setting clear expectations for use.

AI systems can compute, analyze, summarize, and draft-functions that, when guided by instructors, help students learn faster and produce better work. "If you use it to help you be more efficient, to understand something, that can be really time-saving," said Dr. Laura Smith, a senior instructor in the School of Journalism and Mass Communications.

Institutional access and the talent pipeline

USC partnered with OpenAI in June 2025 to provide free access for faculty, staff, and students under a $1.5 million agreement. The goal: give every student hands-on experience with tools they'll meet after graduation.

"AI is expected to change our daily lives in ways we can only imagine today," said Associate Vice President for University Communications Jeff Stensland. "We believe that students need to understand these changes and stay ahead of the curve when it comes to adopting the AI skills that employers will be looking for in new graduates."

How students are actually using AI

Students report using AI for lower-stakes, time-consuming tasks-summaries, outlines, and first-pass drafts. "AI helps me deal with (schoolwork), especially when you don't understand something," said senior Ethan Turkel. "It's very convenient."

Fellow student Caulder Christian sees a trade-off that benefits learning focus: "Giving AI the delegation of the plain stuff gives you more time to focus on what really matters."

The market signal: growth with contradictions

An April report from the United Nations Conference on Trade and Development projected the global AI market could reach $4.8 trillion by 2033, positioning AI as a dominant tech sector by that year. See UNCTAD's AI coverage.

Yet a recent U.S. Census Bureau survey indicates AI adoption among large firms is trending downward, even as usage on campuses expands. Explore Census data on AI adoption.

What concerns educators

Critiques focus on job displacement, privacy, ethics, misinformation, sameness in outputs, and energy use. In education, the core risks are over-reliance and academic integrity.

USC's stance recognizes both sides. "Students need to be responsible about using AI and need to be mindful of academic integrity issues," Stensland said. USC's Office of Student Conduct and Academic Integrity handles code violations related to AI use.

Policy that enables learning without shortcuts

• Define allowed use by assignment. Examples: brainstorming, outlining, clarifying concepts, style suggestions. Prohibit final-answer generation where it defeats learning goals.
• Require transparent disclosure. Ask for a short AI use note (tool, date, purpose) and key prompts. Screenshots or exported chats work when feasible.
• Set citation rules. Include model name/version, date accessed, prompt summary, and URL when applicable.
• Assess the process, not just outputs. Collect drafts, prompt logs, revision histories, and brief oral or written reflections.
• Teach verification. Have students fact-check AI claims, add sources, and flag uncertainties.
• Protect privacy. Prohibit submitting personal, confidential, or proprietary data to public tools. Prefer institution-provided accounts.
• Use detection carefully. Treat any detection score as a signal, not proof. Center decisions on evidence and student work artifacts.
• Promote equity. Ensure all students have access to approved tools or provide alternatives.
• Mind efficiency and energy. Encourage concise prompts, batched queries, and fewer unnecessary regenerations.

High-value classroom uses that work

• Feedback at scale: Generate rubric-aligned comments on drafts, then refine in your voice before sharing.
• Question banks: Draft varied practice items and explanations; validate accuracy before release.
• Reading support: Create summaries at different reading levels, glossaries, and guided notes.
• Language support: Offer translation and grammar help while requiring original ideas and sources.
• Data and code starters: Produce baseline scripts or analysis plans students must adapt and explain.
• Lesson planning: Draft outlines, case studies, and discussion prompts tied to your learning outcomes.

Implementation playbook for departments

• Start small: Pilot in one course or assignment with clear boundaries.
• Share exemplars: Provide examples of acceptable AI use and unacceptable shortcuts.
• Collect feedback: Ask students what helped, what confused, and what should change.
• Align policies: Coordinate across courses to reduce mixed messages for students.
• Iterate each term: Keep what works, cut what doesn't. As Dr. Smith notes, standardization can limit experimentation-this is the window to test and learn.

Bottom line for educators

AI isn't a replacement for teaching. It's a tool that, when directed with clear rules and authentic assessment, saves time and improves learning. Set boundaries, make the process visible, and keep the focus on thinking, evidence, and integrity.

Further resources

• UNCTAD: Artificial Intelligence
• U.S. Census Bureau: AI adoption indicators
• Complete AI Training: Courses by job