Human-like Toy Robots
Decompressing after the Hong Kong Electronics Fair, I started to think about the small robot toys that were being presented there. It occurred that despite the huge gains in AI, natural language, and voice interaction, these human-like robots have evolved at a snail’s pace over the past few years.
There appears to be very little difference between the NAO robot from 2006 an the consumer robots that are on the market in 2016.
That’s not to say that there haven’t been enormous advances in robotics overall in 10 years — from autonomous vehicles coming to market to warehouse robots and drone deliveries. Industrial robots like Baxter and war robots like those of Boston Dynamics and becoming scarily biologic.
However, the market for robot toys hasn’t matured yet and as a result, there are still many robots that are either toy or try to mimic humans that are not compelling. Worst offenders are those trying to be human that end up in the uncanny valley.
Many smaller companies have jumped on the success of the Jibo campaigns and have tried to ship Jibo-like friendly-faced robots that don’t function well. For these robots to be compelling, they need to function at a very low error rate.
That’s simply not the case yet. These robots don’t track well, respond slowly, and have limited applications.
We shouldn’t despair over the current state of consumer robots on the market. This just means they’re primed for a disruption. Over the next 24 months, hopefully with Jibos finally shipping, we’ll see tools that will come out that will make for much more compelling user interactions.
Different tools like Google’s and Microsoft’s computer vision APIs will help get these toy robots past the low value application humps and combined with working voice interaction, these devices will be easier to interact with.
