Education and Technology

an interview with Dr. Kristyn Sommer

Hi! Please tell us a bit about yourself.

Hi! I’m Dr. Kristyn Sommer. A developmental scientist, mum of two, and science communicator who translates the child development research into practical parenting and education insights. I’m autistic and (almost finished with my diagnosis for) ADHD, and I care deeply about making learning and therapy spaces more inclusive, especially for neurodivergent kids. I also spend a lot of time researching how children learn from non-traditional agents like robots and screens. You’ll usually find me on every kind of social media (@drkristynsommer) busting myths about parenting, screens, and the science of how kids actually learn.

The WHO and American Academy of Pediatrics (AAP) suggest toddlers (18‑24 months) get only brief doses of high quality, educational screen time. What should caregivers or teachers look for in an app or platform to meet these suggestions?

Great question! For this age group, quality outpaces quantity every time. Look for apps and shows that are developmentally appropriate (simple, slow-paced, and clear) and interactive (not just passive watching). Bonus points if the app encourages interaction with the real world (e.g., with parents or peers). Any easy way to pick a better form of screen time is to think about it like this: Is the app or show treating your toddler like a partner in learning, or just blasting them with noise and colour?

Past studies found that children under three copy actions better from a live adult than from a prerecorded video—a phenomenon called the “video deficit.” However, your 2023 study didn’t find that gap, and this is supported by a recent meta-analysis indicating that the video deficit is steadily decreasing. What might be causing this change?

The combination of the latest evidence suggests that the “video deficit” may not be a hardwired developmental limitation but instead a moving target, shaped by context and culture. Kids today are growing up in media-saturated environments where video calls, livestreams, and screen-based interaction are part of daily life. So they’re becoming experts at parsing social cues from screens. In other words, screens aren’t “less real” to them like they were for earlier generations, they’re just another medium for communication. It’s no longer a distinction between real life and screens for them, it is all just “life”.

The WHO and AAP guidelines mentioned above encourage parents to watch alongside toddlers. Given evidence indicating the disappearance of the video deficit, would you say that solo screen time can be just as educational?

Short answer: sometimes, but not always. While the video deficit is shrinking, co-viewing still plays a massive role, especially in early toddlerhood. A caregiver sitting next to a child, naming things on the screen, asking questions, or even just being emotionally available helps turn screen time into shared attention, which is where the magic happens. For older toddlers or children, well-designed interactive media can absolutely support independent learning, especially when it builds on what the child already knows. Ultimately, we should always opt for connection and co-viewing, but we certainly don’t need to carry guilt about occasionally asking for a break with the support of a screen if needed.

In your 2024 study, you found that children ages 3–6 displayed a robot bias. Can you explain what this means and its implications for the potential of AI in education?

Absolutely! A “robot bias” means that when given the option, kids sometimes trust or learn more from a robot than a human. It sounds wild, but it reflects how kids evaluate expertise and social relationships. They often see robots (and technology more generally) as reliable, consistent, and nonjudgmental. That has big implications for AI in education. Kids may treat these systems as credible learning partners. So we have to be very careful about what the AI is teaching and how it’s framed.

Schools are piloting AI tutors. What classroom conditions are essential for these tools to ensure safety and improve learning outcomes?

We are starting to approach the limits of my personal expertise, but I can offer some insights. Three things I can see being important:

1. Human oversight – AI should augment, not replace, teachers. Educators need to be able to monitor, intervene, and contextualise learning. I’ve spent a decade researching robotic teaching assistants and the goal has always been to support teachers in doing the important human parts (e.g., connection) while the robots do the heavy lifting of the more monotonous tasks.
2. Transparency – Students (and teachers!) should know how the AI works and what it’s doing with their data. Students should be armed with the tools to evaluate the knowledge the AI gives them.
3. Pedagogical alignment – AI tools must be designed around actual educational goals, not just tech novelty. No one needs another chatbot teaching random facts, a room full of $20,000 robots gathering dust or a stack of iPad rotting in the corner of the classroom.

With increasing technology use in the classroom, how do you see the educators’ roles changing, and what new skills will educators need to work effectively alongside these tools?

Again, this topic is beyond my scope of expertise, so I will share only brief insights and defer to others more expert in this domain than me.

Teachers may need to shift from being information deliverers to learning facilitators, which, to be frank, has already been occurring for many years in some education environments. That means more emphasis on guiding critical thinking, supporting collaboration, and tailoring instruction based on insights from tech tools, rather than standing in front of a class just lecturing content for rote learning.

Educators will need to upskill in their digital literacy, particularly around AI and knowing when a tool is helping and when it’s hindering. Teachers will (hopefully) still be the arbiters of knowledge and that means they will also need the confidence to say no to tech that isn’t developmentally appropriate or evidence-based and draw upon more manual methods were required.

In which learning scenarios do you think AI has the greatest potential to improve student outcomes?

AI shines when it personalises feedback in real time—especially for skills like maths, reading, or even second-language learning. It can also be powerful for students who need repetition, scaffolding, or nonjudgmental practice spaces (like anxious or neurodivergent learners). So I think there are many places where AI has great potential for learners. But the biggest impact that AI will have on learners will likely depend on how it is used alongside their educators. The key will be ensuring that AI tools empower teachers—not sideline them.

To learn more about Dr. Sommer's work, visit her Google Scholar profile or follow her on Instagram, TikTok, or Facebook: @drkristynsommer.

Read other Media Literacy Interviews.