Profile image

Dylan Bulseco, Ph.D.

Senior Vice President & Principal Scientist
Institute for Future Intelligence

Studying how artificial intelligence reshapes pedagogy, scientific practice, and human expertise in education and the life sciences.

My Mission

The integration of artificial intelligence into education represents a pivotal shift, demanding a re-evaluation of our pedagogical values and workforce development strategies. As AI rapidly evolves from machine learning to large language models, its impact on education requires a human-centric approach, prioritizing critical thinking and ethical reasoning along with quantifiable metrics like test scores. "Human-in-the-Loop" must always be considered as we design and deploy AI-enabled technologies. We must actively guard against "goal drift," and ensure that AI tools support deeper understanding, critical thinking, creativity, and civic engagement rather than merely optimizing for measurable outputs. The future workforce requires more than technical proficiency. It also requires the ability to critically evaluate AI-generated information—resulting in a mindset that questions "Why should I trust this output?". This means we must promote AI literacy, scientific communication skills, and a robust understanding of the scientific process itself, moving beyond memorization and emphasizing understanding. Ultimately, our mission in education should be to help students view AI as a tool that can augment, not replace, human intelligence, empowering them to ask the right questions, decide on the the right goals and discern the rules governing its application within science and society. This requires a conscious effort to preserve the uniquely human aspects of teaching – empathy, ethical reasoning, and curiosity – while developing a workforce capable of responsibly engaging with rapidly evolving, sophisticated AI systems.

From a scientific perspective, "goal drift" in the context of AI integration in education refers to the situation where the objectives of AI implementation diverge from the intended, community-valued goals. Specifically, it describes a situation where AI systems are optimized for metrics that are easily quantifiable by the algorithm (e.g., test scores, efficiency) at the expense of broader, less measurable educational outcomes such as fostering creativity, critical thinking, ethical reasoning, and civic engagement. This represents a form of optimization bias, where the system prioritizes what it can measure over what is valued by society as a whole. To prevent goal drift, the provided context suggests a proactive, iterative approach framed as an ongoing ethical and educational inquiry.

Key strategies include:

Professional Background

Dylan Bulseco is a Senior Vice President and Principal Scientist at the Institute for Future Intelligence (IFI). He holds a Ph.D. in Biochemistry and Biophysics from Oregon State University and brings over two decades of experience in fluorescence-based instrumentation, reagents, and analytical applications spanning research, diagnostics, medical devices, and therapeutics.

His career bridges academic research and commercialization, with leadership roles as Principal Investigator, co-Principal Investigator, or senior personnel on federally funded projects supported by the National Institutes of Health (NIH), the National Science Foundation (NSF), and the Department of Homeland Security (DHS).

Focus on AI in Education

At IFI, Dylan’s work centers on studying the impact of artificial intelligence on education, particularly in STEM and life sciences contexts. His research examines how AI tools—such as intelligent tutors, generative models, and data-driven learning systems—transform pedagogy, assessment, and knowledge construction.

Rather than treating AI as a replacement for educators or expertise, his work emphasizes AI as a pedagogical partner: a tool that augments human reasoning, supports inquiry-based learning, and enables educators to design more adaptive, reflective, and equitable learning environments.

This research draws on transcript analysis, learning sciences theory, and human–AI interaction studies to understand not only what students learn, but how learning unfolds through dialogue, reflection, and guided exploration in AI-enhanced educational ecosystems.

Cultural Relevance and Workforce Development

As a Native Hawaiian, Dylan is deeply committed to culturally responsive science education and workforce development. His projects integrate computational science, data science, and artificial intelligence into life sciences education in ways that respect local knowledge systems and community contexts.

He has been actively involved in initiatives supporting science education and workforce pathways in both Massachusetts and Hawaiʻi, with a particular focus on community colleges, place-based learning, and experiential research programs that prepare students for emerging STEM careers.

Source for Material

Although development of all ideas and blogs for this site are mine, an AI-enabled workflow is always used throughout the development process. A custom Research Agent—built to run local large language models via Ollama—is used to investigate topics, perform gap analyses, conduct targeted web searches, and generate initial outlines or draft posts; its source code is available on GitHub. Most images are produced using Google Notebook. Draft blog edits require a mix of hands-on review and revision as well as the use of a variety of LLMs, including ChatGPT, Gemini, and locally deployed models such as gemma3:27b, gemma3:12b, and gpt-oss:latest. Final posts are reviewed then prepared by uploading a formatted .docx file into a lightweight pipeline that converts it to HTML, enters relevant information into a local MySQL database, and generates static html pages for hosting on Eco Web Hosting, using shared (inexpensive) hosting services. The site-generation code is also publicly available in my GitHub repository.

Learn More