Sarah Austin is a researcher, educator, and SDG advocate working at the intersection of education, artificial intelligence, and human centered technology design (HCI). A former Visiting Fellow at Stanford University and recipient of the University of Utah Faculty Award, she currently works alongside teachers and students on applied educational research focused on learning, cognition, and educational innovation. Austin is Co-Founder of Tilted AI, which supports research into HCI and student scholarship advancement. She has also been recognized by the United Nations as a Humanitarian Fellow in Artificial Intelligence and serves as an advisor and chairwoman to Model United Nations organizations and student leadership initiatives worldwide.
For more than a decade, educational technology has been defined by a simple assumption: more access to digital tools will create better learning outcomes. Classrooms have become increasingly connected, students have gained access to unprecedented amounts of information, and schools have embraced platforms designed to personalize instruction, increase engagement, and improve educational equity. Yet as digital learning environments become more deeply integrated into everyday education, a new question is beginning to emerge among educators, researchers, parents, and technology leaders alike. How can schools steward technology to support human cognition rather than competing with it?
This question sits at the center of a growing conversation surrounding student attention, mental fatigue, memory retention, and the long term developmental impact of increasingly screen intensive learning environments. Teachers across grade levels frequently describe students who appear constantly connected but increasingly distracted. Many report observing difficulties with sustained concentration, deep reading, information retention, and the ability to engage in focused cognitive effort for extended periods of time. While these challenges cannot be attributed to any single cause, they’ve prompted researchers to examine how digital environments influence the neuroscience behind how students process information, manage attention, and learn.
As artificial intelligence accelerates the evolution of educational technology, researchers are increasingly exploring whether the next generation of learning systems should be designed not simply for engagement, but for what some are beginning to describe as cognitive sustainability.
Educational technology has spent much of the last decade optimizing for engagement. The next decade may require optimizing brain cognitive sustainability. While engagement remains important, startup companies like Tilted, a space for AI learning and agentic loop management, are sponsoring research into educational tech interfaces that encourage sustained attention, meaningful reflection, memory formation, emotional regulation, and deeper forms of learning. EdTech success may no longer be measured solely by how frequently students interact with screens, but by how effectively AI supports the cognitive processes that underpin long term learning.
For decades, educational innovation has often been evaluated according to what technology can do. Increasingly, researchers are asking a different question: what does technology do to the learner? Beyond academic performance metrics and completion rates, attention is turning toward cognitive resilience, executive function, emotional wellbeing, and the developmental effects of growing up in highly digital environments. This shift does not reject innovation. Instead, it reflects a broader effort to ensure that EdTech evolves in ways that align with the needs of developing minds.
Neuroscience research offers important context for this discussion. Human attention is a finite resource. The brain continuously filters information, prioritizes stimuli, and determines what is worthy of deeper processing and memory storage. Learning is not simply the act of encountering information. When attention becomes fragmented, learning itself can become fragmented.
With continuous updates, notifications, and frequent shifts in focus screen time becomes addictive. While these design principles can increase short term engagement, researchers have increasingly examined if constant stimulation may contribute to cognitive overload. Students frequently move between messaging platforms, educational applications, videos, social media environments, search tools, games, and AI systems throughout the day. The cumulative effect of these transitions may influence attention management, mental fatigue, and the ability to sustain focus on complex tasks like research.
Students may appear mentally exhausted despite spending much of their day consuming information. They may struggle to engage in reading, maintain concentration during lectures, or retain information. Teachers report observing “brain fog” a state characterized by cognitive fatigue. While research sponsored by Tilted looks at the mechanisms underlying these experiences, interfaces must do more than deliver information. The research has found that AI can’t be the primary source of school work because it risks turning students into consumers of content. Rather interfaces must require that students create the work on their own first and only use AI to enhance their work or to revise their work. Tilted’s research has discovered that when AI is used in revisionist student workflows it encourages dialogue, inquiry, and active participation that can strengthen the very cognitive processes that support long term learning.
The research adopts an applied educational research framework grounded in authentic classroom practice rather than simulated learning environments. Through longitudinal observation of students and teachers operating within live instructional settings, the study investigates how voice first AI systems and conversational learning interfaces influence learner engagement, cognitive load, verbal language development, classroom participation, and instructional effectiveness. Researchers are collecting both qualitative and quantitative evidence through classroom observation protocols, educator interviews, student feedback, learning analytics, and literacy focused assessment measures. The project is informed by ethical educational technology principles and Sustainable Development Goal 4 (Quality Education), with particular attention given to educational equity, accessibility, teacher augmentation, and the responsible deployment of AI within underserved learning communities. By examining technology within the social and cultural realities of classrooms, the research aims to generate practical evidence capable of informing future educational policy, technology design, and pedagogical practice.
This distinction may become increasingly important as schools adopt AI. Emerging AI systems, however, are beginning to support more conversational forms of learning. Rather than requiring students to navigate screens by menus, click through modules, or consume information passively, agentic conversational AI can engage learners through dialogue, questioning, explanation, and verbal reasoning. In many respects, these interactions resemble some of humanity’s oldest learning tradition, auditory learning between teachers and students.
The growing interest in voice first learning environments reflects this shift. Advances in speech recognition, natural language processing, conversational AI, and agentic systems are making it possible for students to interact with educational technology through spoken language rather than exclusively through visual interfaces. Instead of relying on constant screen interaction, learners can increasingly ask questions, receive explanations, practice language skills, engage in verbal problem solving, and participate in guided educational experiences through conversation.
This development is particularly interesting because human learning has historically been rooted in oral communication. Long before textbooks, tablets, or Google Classroom existed, knowledge was transmitted through storytelling, dialogue, mentorship, apprenticeship, and rhetoric. Children learn language through listening, debating peers, and speaking. They develop understanding through conversation.
Researchers are also beginning to examine whether voice first interactions may help reduce certain forms of cognitive load associated with screen intensive learning environments. High visual digital experiences can dominate cognitive attention while reducing opportunities for auditory processing, listening comprehension, and verbal communication. Although the relationship between screen use and developmental outcomes remains complex, these findings have encouraged educators to explore learning models that place greater emphasis on speaking, listening, conversation, and oral language development alongside traditional literacy instruction.
This is particularly relevant in regions where access to qualified educators remains limited. Across many developing regions, including parts of rural country or isolated urban environments, schools may not exist or the ones that do continue to face significant shortages of trained instructors. English language acquisition remains a critical challenge for millions of learners. Emerging conversational AI systems such as Elsa Speak help address some of these gaps by providing students with opportunities to practice pronunciation, phonics, listening comprehension, and conversational language skills through interactive voice based experiences. While such systems cannot replace teachers, they may expand access to individualized practice and support.
Another concept gaining traction within both education and technology is ambient computing. Unlike traditional software that demands continuous user attention, ambient technologies are designed to operate in the background, becoming available when needed without dominating the user’s experience. In educational settings, this philosophy aligns closely with the goals of cognitive sustainability. Rather than requiring students to remain continuously engaged with screens, ambient systems can support learning while preserving attention for teachers, peers, discussion, and original work.
This broader movement reflects an important philosophical shift. For years, educational technology innovation often focused on increasing screen interaction. Today, some of the most interesting developments seek to reduce unnecessary interaction altogether. The objective is not to remove technology from learning environments but to make technology less intrusive, more natural, and more aligned with the ways students learn.
Drawing from neuroscience, human computer interaction, developmental psychology, language acquisition research, and AI, Tilted’s sponsored research aims to better understand how educational systems can support attention, engagement, literacy development, and learner wellbeing in congruence with the United Nations Sustainable Development Goals.
The significance of this research extends beyond any single technology platform. At its core, the work reflects a broader question facing educators globally. What should educational technology look like in an age of AI? If AI becomes a permanent part of education then how can schools ensure that it strengthens human capabilities rather than diminishing them? How can technology encourage curiosity rather than distraction, participation rather than passivity, and understanding rather than information overload?
The answers may ultimately lie not in eliminating screens, nor in embracing them uncritically, but in designing learning experiences that respect the realities of human cognition and original work because getting answers wrong is part of the learning process. Failing is part of learning. The future of educational technology may be less about maximizing screen time and more about maximizing meaningful peer to peer interaction. It’ll likely be less about delivering information and more about cultivating debate.
As generative AI becomes increasingly capable, schools may need to place less emphasis on evaluating what students submit and greater emphasis on evaluating what students can explain, defend, and apply. While AI can cheat with homework, essays, and written assignments, it cannot easily replace a student’s ability to engage in debate, articulate reasoning, respond to challenges, or defend a position in class. This is driving renewed interest in assessment models rooted in oral examination, rhetoric, presentations, seminars, and programs such as Model United Nations, where understanding must be demonstrated through live discussion rather than simply handed in as written work.
The future of assessment may be less about what students hand in and more about what they can defend.