My career in AI began in the White House, where I worked for the U.S. chief technology Officer and later as the chief data officer of USAID. We used technologies that laid the groundwork for AI, opening up vast amounts of government data to improve crisis response efforts.
Over time, it became clear that machine learning—or AI models, as they’re now called—could significantly outperform traditional statistical models in predicting high-impact, low-frequency events like those exacerbated by climate change. This led me to co-found Kettle, an insurance startup that develops AI models trained to better predict wildfires and other crises caused by climate change.
My path has been driven by a passion for harnessing technology to solve complex problems. Still, it has also made me acutely aware of AI’s limitations, particularly in the nuanced world of education.
From White House to Wayfinder
Before co-founding Kettle, I worked closely with Patrick Cook-Deegan, founder and CEO of Wayfinder, to build a program that would give K12 educators tools to help students develop social-emotional skills to thrive in and out of school, cultivating lives of belonging and purpose.
As an organization concerned with young people’s social-emotional development, Wayfinder didn’t initially have AI on its radar. However, as part of its commitment to guiding students toward future readiness, it must now help educators grapple with the reality of teaching in a world where students can access AI and a future in which they may have to compete and collaborate with AI tools.
Throughout my career, I’ve seen that AI excels at solving specific problems, like safe driving or answering factual questions, but struggles with dynamic systems requiring adaptability, such as classrooms or businesses. While concerns about AI are valid, certain human aspects of education and work will remain irreplaceable.
What educators need to know about AI’s limits and possibilities
Educators have always had to adapt to technological advancement, and I see AI as the next evolution of this process. Consider the current state of AI in classrooms akin to the advent of Google: When Google first appeared, there were concerns that it would undermine academic research. Instead, it made tasks like finding sources more efficient, allowing more time for creativity and deeper inquiry. Similarly, AI should be viewed as a tool that can augment human capabilities, not replace them.
‘Talking Out of School’ podcast: Why media literacy now matters so much
Teachers can harness AI’s capabilities to rethink how they teach and assess skills without overturning their practice. For example, if AI can write essays, how will teachers assess students’ evidence-based reasoning? They could consider having students analyze multiple arguments and make a case for which is soundest. This would change student output, but it could also help them develop similar critical thinking skills and learn various persuasive techniques.
Another practical use of AI in the classroom would be to assign students the task of determining what is AI-generated. Honing this critical capability may be a crucial skill in the future and students can start building it now as part of their digital literacy development.
As a tool for teachers, it can help make lessons more individualized, accessible and engaging. It can reduce teachers’ workloads, helping improve conditions that contribute to burnout. For students, it can automate rote learning tasks and create more opportunities to develop higher-order thinking skills.
Here are just a few ways AI tools can be incorporated into K-12 instruction in coming years:
- Research assistance: AI-driven research tools like Consensus can help students quickly find relevant and peer-reviewed research. With less time needed for finding and validating research, students could focus on synthesizing information, developing evidence-based arguments and reaching innovative conclusions.
- Language support: Educators can leverage AI to provide language scaffolds for multilingual learners, making learning materials accessible to a broader range of students. Because tools like ChatGPT operate on language models, they are well-suited to this particular task.
- Automated grading: AI-assisted grading will save time for teachers, allowing them to focus on planning interactive and personalized instruction to better nurture students’ strengths and help them develop in areas of need.
- Formative assessments: Real-time assessments that can offer immediate feedback about students’ skill development could help educators tailor their teaching strategies to the needs of their classrooms.
- Adaptive learning platforms: Currently used more commonly in professional settings, adaptive learning platforms can adjust the difficulty of tasks based on a learner’s performance. In the classroom, such platforms could provide more personalized learning experiences, helping teachers better serve classrooms with diverse learning needs.
- Interactive simulations: As AI becomes more widely available and easier to use, it could offer numerous opportunities for simulations and virtual labs to provide learning experiences otherwise impossible in most K-12 classrooms.
With AI’s possibilities come responsibilities. Educators must teach students strategies for the effective and responsible use of AI tools. For instance, ChatGPT can mislead students who trust it like a search engine. Teachers need to help students recognize biases in AI algorithms, evaluate AI-generated information and consider ethical implications. By fostering these skills, educators prepare students to navigate a future where AI is prevalent, enabling them to use it effectively while understanding its limitations.
We can expect a future workforce characterized by career paths previously considered unconventional. With the ever-quickening rate of technological advancement, it’s essential to remember that the skills students develop through human interactions—like self-awareness, collaboration and agency—are irreplaceable and enduring.
Moving forward with AI as a co-pilot
With early studies showing the negative impact of AI-driven virtual tutoring, it’s clear that technology will never recreate the critical, relational aspects of education. Instead, it can offer teachers a helping hand, making their jobs more manageable and their instruction more effective.
As we continue to guide students to develop future-ready skills, our goal should be to empower them to be adaptable, empathetic and purpose-driven individuals who can thrive alongside evolving technologies. By understanding and embracing AI’s possibilities and limitations, we can ensure that students are prepared to use AI effectively and lead in a world where technology and humanity intersect.
