AI Toys and Child Safety

From ELIZA to ChatGPT, conversational generative AI has been a part of our world since 1966. Inevitably, it made its way into the the world of toys and play. In recent years, a number of notable toys; My Friend Cayla, Hello Barbie, and ROYBI Robot just to name a few, touted conversational AI as a primary selling point. But few of these were successful, most found themselves discontinued after a short production run, and a handful even landed their manufacturer in hot water.

As more advanced generative AI is developed, it’s likely that more toys that use this tech will hit the shelves. It’s imperative that consumers are well informed on their threats and promises while designers are taking these risks and opportunities into account in their practice.

Threats

Conversational generative AI toys collect audio and video recordings in order to adjust their behavior and better converse with a child. This can have concerning implications for safety and privacy if manufacturers of the toy aren’t transparent about how they use and store that data. These toys are often also connected to the internet making them susceptible to malicious remote control.

Additionally, as you might already know, AI driven tools are prone to algorithmic bias. If the data fed into a tool is not varied enough, it won’t be able to adequately support the unique characteristics of a child, leading it to potentially cause more harm than good.

Beyond these more obvious threats, there’s one lurking in the shrouds of a feature and that is the false promise that these toys can stand in for or even replace invaluable child-parent interactions. There are of course children whose parents are unable to be fully present for one reason or another and maybe in those situations, the conversation provided by an AI is better than nothing at all. But nothing beats human-to-human, face-to-face time with parent figures.

Promises

On the bright side though, this technology really can support children and their parents in ways that enhance child development and family relationships. Play time can be a deeply enriching experience and these toys can nurture a child’s unique strengths and interests by adjusting their behavior patterns. CALL theory suggests that these toys have the potential to support language learning while the implicit network int these toys could help create a personal learning environment for children.

These toys could also record data about a child’s habits with the intent of helping them build healthy habits or even diagnosing illnesses in early stages.

Conclusion

In order to insure that products targeted at kids are safe, we need more regulations akin to COPPA and we need systems in place that allow the FTC and other relevant regulator authorities to check the safety of toys well before they hit the market. We also need to delegate the responsibility of checking for algorithmic bias to a new or existing authority and create a culture of research informed toy design.

Well designed technology can help us build a better world and its up to you, me, and us to stay critical of innovations and demand responsible development. In our techno-centric world, we often have a tendency to talk about technology only through the lens of what it gives us. But it’s crucial that we never forget to ask what it takes away.