The promise of Artificial Intelligence often centers on a single, seductive idea: automation. We envision a future where AI handles the mundane, repetitive, and time-consuming tasks, freeing us up for… well, for what exactly?

While the vision of a “universal teammate” dutifully checking off our to-do list is appealing, it only tells half the story. The other, more profound half, is about augmentation—using AI not just to do things for us, but to help us think better.

This was the core theme I took away from the ASU conference on Agentic AI and the Student Experience, where Danielle Perszyk, a Cognitive Scientist at Amazon's AGI Lab challenged us to shift from a purely efficiency-based view of AI to one focused on deeper human-computer symbiosis.

The key to answering this question lies in a concept that might seem counterintuitive: friction.

We’ve been conditioned to believe that all friction is bad. In the world of user experience, the goal is almost always to create a “frictionless” journey. And for many tasks, that’s exactly what we need.

When AI automates repetitive tasks, it effectively reduces friction. This is the foundation of curiosity and exploration; when the cost of trying something new is low, we are more likely to do it.

However, there’s a hidden danger in a completely frictionless world. When our tools do all the work, our own skills can atrophy. Think of the auto-complete feature in your email. It’s undeniably efficient, but it can also make us lazy writers.

This is where the other side of the coin comes into play: increased friction. Not to create frustration, but to introduce a healthy dose of “cognitive friction.” This is the kind of friction that prompts us to pause, to think more deeply, and to engage with a problem on a more sophisticated level.

It’s the difference between being given the answer and being guided to discover it for yourself.

The history of technology is a story of “techno-social co-evolution.” We invent tools, and then those tools shape us. The printing press didn’t just make it easier to copy books; it fundamentally changed how we share and develop knowledge. The same will be true of AI.

The real revolution won’t be in the tasks we automate away, but in the new ways of thinking we augment.

This is where the concept of a "universal teammate" becomes particularly interesting. Imagine if this teammate not only handled your expense reports but also served as a sparring partner for your ideas. It could challenge your assumptions, point out blind spots, and present you with conflicting information to help you build a more robust argument. This is the essence of augmentation.

Perszyk emphasized that sophisticated thinking is often latent until we interact with others. General intelligence, she argued, emerges through our interactions—which means we should be measuring the quality of our interactions with AI agents, not just their output. A transcript of interaction can reveal the "moment of insight," allowing us to distinguish between deeper learning and surface-level memorization. This has profound implications for how we design AI systems, particularly in educational contexts where every student could potentially have a team of agents supporting their learning journey.

The path to a future where AI makes us smarter is not about building more powerful automation. It's about building more thoughtful augmentation. It's about embracing the right kind of friction, the kind that challenges us to be better. It's about designing AI that acts not just as a servant, but as a true cognitive partner—what Perszyk calls "tools for thought."

The next time you think about the future of AI, I encourage you to look beyond the allure of automation. Ask yourself:

Florida State University (FSU) is focusing on faculty empowerment through AI by providing low-code/no-code tools that allow instructors to create and manage their own AI agents for classroom use.

Key insight: The process should be frictionless; faculty don’t want extra administrative work to set up or maintain agents.

Across universities and colleges, leaders are reframing what scale means in the age of AI.

Victoria Maloy from Iowa noted how the adoption of Gradescope began in one department but spread organically through faculty networks, demonstrating that sustainable innovation grows from peer sharing and grassroots advocacy rather than top-down mandates.

Feng Hou, CIO of St. Louis Community College, shared how AI-powered tutors, grant-writing agents, and virtual assistants have reduced administrative load and extended 24x7 support for students — allowing faculty and staff to focus on teaching, mentoring, and problem-solving.

Kemi Jona from the University of Virginia reinforced this theme, noting that when the University used AI to generate course videos, they cut costs by half and reclaimed creative time.

Even as optimism grows, leaders remain clear-eyed about the challenges. Donna Kidwell, CIO of the University of Toronto, highlighted the complexity of managing privacy in a world where students are experimenting with multiple AI tools — from GPT to Character AI. She highlighted that responsible innovation must balance creativity with compliance, especially in countries like Canada with strong privacy regimes.

Cost and integration barriers persist as well, a concern echoed by Maloy and others who view sustainable AI adoption as dependent on financial and infrastructural agility.

In a world where careers transform at unprecedented pace, the summit emphasized durable, portable skills over job-specific training. This reality demands a fundamental shift in educational philosophy. Rather than preparing students for specific positions, institutions must equip them with enduring capabilities that transcend individual jobs and industries.

Durable skills are fundamental human capabilities that remain valuable and relevant across various roles, industries, and technological advancements. They are not easily automated by AI and form the bedrock of complex problem-solving and interpersonal effectiveness.

Examples include: Critical thinking, creativity, complex problem-solving, communication, collaboration, adaptability, emotional intelligence, and ethical reasoning.

In the age of AI, these skills become paramount. As AI automates routine and analytical tasks, the demand for uniquely human attributes — like innovative thought, nuanced communication, and the ability to navigate ambiguity — intensifies. Cultivating durable skills prepares students to lead, innovate, and thrive in an AI-augmented workforce, where their value lies in contributions that AI cannot replicate.

Portable skills are competencies that can be transferred and applied effectively from one job, industry, or context to another. They act as bridges, enabling individuals to pivot their careers, adapt to new organizational structures, and leverage their expertise in diverse environments.

Examples include: Project management, data analysis, digital literacy, research methodology, strategic planning, negotiation, and public speaking.

For students and educators, understanding portable skills is crucial for fostering career resilience. As industries merge and new roles emerge with AI integration, the ability to transfer skills across domains ensures continuous employability. Educators should focus on teaching the underlying principles of these skills, rather than just their application within a single context, preparing students for dynamic career trajectories.

Current skills refer to specific technical proficiencies and specialized knowledge that are highly in-demand in the immediate job market. These skills are often tied to specific technologies, software, or industry trends and have a relatively shorter shelf-life due to rapid innovation.

Examples include: Proficiency in specific AI frameworks (e.g., TensorFlow, PyTorch), cloud computing platforms (e.g., AWS, Azure), specific programming languages, cybersecurity tools, or advanced data visualization software.

The ASU Agentic AI summit painted a compelling and transformative vision for higher education's future, one deeply rooted in human potential and technological empowerment.

In this future, AI transitions from being a simple tool to an integral, intelligent companion throughout the learning process. It serves as a personal tutor, offering real-time, adaptive feedback on complex assignments, guiding students through challenging concepts at their own pace, and recommending resources tailored to individual learning styles and needs. This adaptive approach ensures that no student is left behind, and every student is challenged appropriately, fostering deeper engagement and mastery.

Concurrently, the summit emphasized empowering faculty to be the architects of this AI-driven evolution. By providing accessible, intuitive AI tools, educators can streamline administrative tasks, freeing up valuable time to focus on mentorship, research, and innovative teaching strategies.

Imagine AI assisting in generating diverse quiz questions, summarizing vast amounts of research for curriculum development, or even helping design engaging, interactive course content. This empowers faculty to experiment with new pedagogical models, develop highly personalized learning paths, and ultimately, elevate the quality and relevance of their instruction without being burdened by technological complexities.

Practical examples abound: AI can suggest project topics that align with a student's emerging career interests, provide virtual labs and simulations for hands-on experience, or facilitate collaborative learning environments with intelligent group formation. This ensures content is not just consumed but actively applied, allowing students to develop critical thinking, problem-solving, and adaptive skills that transcend specific jobs and prepare them for a dynamic, AI-augmented workforce. The goal is to cultivate a generation of learners who are not just users of AI, but creators, innovators, and ethical leaders in an evolving technological world.