Apple has a Machine Learning Research department, and it's yielded ELEGNT, a robotic desk lamp clearly inspired by Pixar. The lamp can be operated by both voice and gesture, and the researchers gave it two modes of doing things: "Expressive," and "functional."

In the expressive mode it acts, well, cute, borrowing body language from dogs in order "to interact more naturally with humans. In the functional mode, the lamp simply does what you ask it to do (change the direction of the lighting, for instance) without the theatrics. This will make more sense once you see the demo video:

Enter a caption (optional)

I like the gesture control and tracking ability, and the projected homework assistance is an interesting feature. But if I owned one of these as a product, I'd only ever set it to functional.

There's a reason for that. Previously, we looked at Disney Research's work in designing cute, expressive robots. Their goal was also to enhance human-robot interactions.

Last year I felt this was a promising direction of research, but I've rethought that position, giving the way things are going in the world. If right now you were to log onto any forum—discussing politics, truck accessories, gardening techniques, sci-fi books, dog training, whatever—and write a post disagreeing with or questioning a popular opinion, watch how quickly it devolves. It is very easy to get humans to behave rudely. Creating a future where robots are always nice, lulling people to form attachments to them while human-to-human interactions continue to unravel, may not be wise.