Souvick Ghosh, Ph.D., School of Information

Dr. Souvick Ghosh, assistant professor at the San José State University School of Information, has been selected as the university’s inaugural AI Faculty Fellow for 2025–2026. The appointment marks a significant milestone in exploring how artificial intelligence impacts teaching, learning, and research across the Silicon Valley campus.

Ghosh reflected on the meaning of the honor. “Being named the inaugural AI Faculty Fellow is, first and foremost, a position of responsibility,” he said. “It gives me a platform to reach a broader audience across the school and university—students, faculty, and staff—and to help kick-start AI literacy and outreach efforts that are practical, transparent, and human-centered.”

He explained that the position connects closely with his ongoing work in AI. “The role aligns with my background in conversational systems, applied AI, and ethics, and builds directly on work we’ve already begun: campus AI literacy efforts, public-facing webinars, and creating a human-centered AI minor and certificate for non-technical learners,” Ghosh said.

When discussing his approach to the fellowship, Ghosh described two guiding principles. “I approach this fellowship with two complementary aims: demystifying and safeguarding,” he said. “On the one hand, I want students and instructors to feel confident using AI for appropriate tasks, such as brainstorming, data cleaning, and formative feedback, while knowing when and how to disclose their use.”

He also emphasized the need for ethical consistency and a shared understanding. “On the other hand, I want us to establish shared norms regarding transparency, bias, fairness, privacy, and accessibility,” Ghosh said. “Specifically, this entails offering an AI literacy course for students and faculty, hosting public talks and hands-on sessions, establishing a reading group, and creating reusable materials that enable individuals to apply AI responsibly within their respective disciplines.”

Ghosh highlighted inclusivity as a cornerstone of his work. “A key emphasis is lowering barriers for non-coders,” he said. “For example, my ‘vibe coding with AI’ session shows how thoughtful prompting and iteration can help people prototype and learn without an extensive programming background.”

He outlined several teaching and engagement goals for the fellowship year. “I will continue to develop and deliver Canvas-based AI literacy courses through the AI Faculty Fellows initiative, targeting students and faculty, with a focus on responsible AI use,” Ghosh said. “This includes understanding what constitutes appropriate behavior, where disclosure is expected, and how to weigh benefits and risks with clear reasoning.”

Collaboration will be a central component of these efforts. Ghosh explained, “We will host and record a series of talks and workshops on topics such as building library chatbots and integrating AI into teaching.” He added that his October 30 session, “Vibe Coding with AI,” served as a hands-on experience for non-technical participants to experiment with large language models and learn “what works, what fails, and why.”

Sharing practical tools and resources is another part of his plan. “We are also striving to organize a reading group and develop concise micro-resources, such as short guides and checklists that enable instructors to adopt responsible AI patterns quickly,” Ghosh said. “I have also been presenting at faculty technology shares (T3) at the School of Information to demonstrate AI use cases in our teaching and research, such as generating course podcasts from readings, and supporting colleagues as they adopt these practices.”

Ghosh intends to enhance academic paths that promote ethical AI learning. “In parallel, I will strengthen our curricular pathways by supporting the Human-Centered AI Certificate and the proposed minor, ensuring both emphasize ethics, inclusion, and real-world applicability for learners,” he said.

His ongoing research, which underpins many of these initiatives, examines the interaction between humans and large language models. “I will evaluate LLM-assisted qualitative coding on sensitive topics, such as palliative care on TikTok, to identify where models falter, specifically, in handling emotions, sarcasm, and coping humor and when human judgment is indispensable,” Ghosh said. “I will examine the reliability and fairness of AI-plagiarism detection to prevent harm from false positives and to design better integrity practices.”

He also evaluates how identity factors influence the output of AI-generated content. “I am studying how persona attributes, such as race and ethnicity, shape an LLM’s generated outputs, with the goal of surfacing and mitigating stereotype leakage in AI tools,” Ghosh said.

Looking ahead, Ghosh summarized the overarching vision that connects his teaching, outreach, and research. “The goal is straightforward: to make AI useful and comprehensible to our community, and to ensure its use is accountable to our values,” he emphasized.