New Technology / Ai Development
Track AI development, model progress, product releases, infrastructure shifts and strategic technology signals across the artificial intelligence sector.
AI+Education Summit 2026: Scaling Human-Centered AI – What It Takes to Transform Learning for All
Topic
Human-Centered AI in Education
Key insights
- AI and education has moved from possibility to reality
- The real question is how AI scales in learning and who benefits
- Access to technology in education has scaled dramatically, but learning outcomes remain stagnant
- Underserved students are experiencing worse trend lines in learning outcomes
- AI tutoring studies show potential gains of two to four months of additional learning
- Current AI tutoring studies are early stage with limited samples and short time horizons
Perspectives
Analysis of a panel discussion on human-centered AI in education.
Proponents of Human-Centered AI
- Emphasizes the need for intentional design in AI tools to prevent widening educational gaps
- Highlights the importance of involving diverse stakeholders in the design process
- Argues that AI can enhance educational outcomes if implemented effectively
Skeptics of AI Integration
- Questions the scalability of AI solutions without addressing systemic issues
- Raises concerns about the potential negative societal impacts of AI in education
Neutral / Shared
- Acknowledges the rapid adoption of AI technologies in education
- Recognizes the need for robust evaluation methods to measure AI effectiveness
- Notes the importance of equipping individuals with knowledge and skills to use AI tools responsibly
Metrics
learning_gain
two to four months
additional learning from AI tutoring studies
Indicates potential for AI to enhance learning outcomes if effectively implemented.
some of the recent trials we've seen have reported gains close to two to four months of additional learning.
learning_outcomes
early 2000 levels
current state of learning outcomes
Demonstrates stagnation in educational progress despite technological advancements.
learning outcomes are still at early 2000 levels.
implementation_challenges
too many pilots actually and still not enough implementations
the current state of AI product implementation
This indicates a critical gap between product development and actual usage, which could hinder educational advancements.
the bottleneck is we have too many pilots actually and still not enough implementations
procurement_problems
there's a lot of waste
issues with adopting IT products in education
This highlights the risk of ineffective spending and resource allocation in educational technology.
the history of governments and school districts about adopting IT products is not great. There's a lot of waste.
mental_health_usage
more using AI for mental health and well-being than for schoolwork
shift in AI usage among students
This trend indicates a potential gap in educational support and highlights the need for addressing mental health.
they're more using AI for mental health and well-being than they are for schoolwork.
impact_measurement
it's very hard to provide incentives for things you can't measure
challenges in measuring educational impact
Without clear metrics, the effectiveness of AI tools in education remains uncertain.
it's very hard to provide incentives for things you can't measure
decision_speed
you want to be able to fix it tomorrow
the need for rapid decision-making in education technology
Fast decision-making is crucial for adapting educational tools to user needs.
you want to be able to fix it tomorrow
digital_public_goods
the Gates Foundation has been a pioneer
role of the Gates Foundation in identifying digital public goods
Identifying digital public goods can significantly impact educational ecosystems.
the Gates Foundation has been a pioneer
Key entities
Timeline highlights
00:00–05:00
AI's integration into education is advancing from theoretical possibilities to practical applications, raising questions about its scalability and beneficiaries. Despite increased access to technology, learning outcomes for underserved students remain stagnant, highlighting the need for intentional design in AI tools.
- AI and education has moved from possibility to reality
- The real question is how AI scales in learning and who benefits
- Access to technology in education has scaled dramatically, but learning outcomes remain stagnant
- Underserved students are experiencing worse trend lines in learning outcomes
- AI tutoring studies show potential gains of two to four months of additional learning
- Current AI tutoring studies are early stage with limited samples and short time horizons
05:00–10:00
AI is democratizing product creation, enabling non-technical individuals and those in developing countries to realize their ideas. However, the challenge has shifted from engineering talent to the effective implementation of AI products, particularly in education.
- This is an amazing moment for AI as it has democratized the ability to create products
- Non-technical people and those in developing countries can now make their ideas a reality
- The bottleneck is shifting from engineering talent to the implementation of AI products
- A product has no value unless it is implemented and used
- Human-centered AI is needed to understand user needs and context
- There are procurement and incentive problems in adopting IT products in education
10:00–15:00
The discussion emphasizes the need for human-centered AI design in education, which considers the broader community impacted by AI systems, including teachers, students, and families. It highlights the importance of involving diverse stakeholders in the design process to ensure positive societal effects.
- We need to go beyond user centered design to human centered AI design
- Design must consider the broader community impacted by a system
- In education, design should account for teachers, students, families, and other stakeholders
- Successful AI systems can lead to societal level effects
- AI in social media has had significant societal impacts
- Human-centered AI requires a diverse design team from the start
15:00–20:00
AI is being integrated into education to assist teachers in personalizing learning experiences for students. Tools like New Zella and KIDM are examples of how AI can enhance instructional methods and provide insights into student learning trends.
- AI helps teachers with scaffolding and appropriate reading levels for students
- New Zella has been used by over 4 million teachers
- KIDM integrates instruction assessment and curriculum into one platform
- KIDM allows teachers to see learning trends across the classroom
- AI is disrupting traditional education and human skills
- Almost half of all ChatGPT users are under 25
20:00–25:00
Mental health and well-being are increasingly being offloaded to AI tools, raising concerns about belief offloading and its implications for education and interpersonal interactions. The need for equipping individuals with the knowledge and skills to effectively use AI tools is emphasized, alongside the challenges of measuring impact in educational settings.
- Mental health and well-being are being offloaded
- There is a concern about belief offloading with AI tools
- The choices made today will have outsized effects on education and interactions
- Equipping people with knowledge and skills to use AI tools is crucial
- Decisions will need to be made about when not to use AI
- Measuring impact in education is challenging due to the risk of teaching to the test
25:00–30:00
The discussion centers on the importance of investing in tools to measure AI performance in education, emphasizing the need for robust evaluation methods. It highlights the gap between engagement metrics and actual educational outcomes, advocating for collaboration with third-party groups to establish credibility in educational tools.
- Investing in tools to understand AI performance is crucial for good AI products
- Using students and teachers as guinea pigs in education requires robust measurement and evaluation
- Collaboration with third-party groups is important for establishing credibility and trust in educational tools
- There is often a gap between usage/engagement and actual outcomes in educational technology
- Evidence-driven fact patterns are essential before making claims about educational tools
- Supporting teachers with AI should focus on augmentation rather than replacement