Picking up from our conversation of two weeks ago, we’re continuing to look at the robot Sophia, who is designed to be human-like in facial expressions and eye contact. Sophia can converse easily with you in a natural-speaking voice. She’s a far cry from the stiff and stern robots (with their monotonic voices) that have been portrayed in many (in-the-past) movies and TV shows.

At the same time, while the YouTube interviews with Sophia give us the sense that she can understand speech, and respond intelligently, those interview clips are selected to (mostly) showcase her ability to seem human-like.

The reality? She’s not quite there. Not yet.

Sophia the robot is part Chatty-Cathy (updated for 2020), and part real AI. The important thing is: a lot of what she says is scripted.

Check out this Sophia vid.

I bet that if you go through all of her statements, you’ll realize that just about everything in this YouTube interview is pre-scripted. (Including the famous “OK, I will destroy humans” statement. PLEASE, let that be pre-scripted!)

To get a more realistic sense of Sophia’s current capabilities, check out this vid:

After you’ve done a personal reality check, read this article in : The Verge

So … I’m going to postpone a little “where’s the AI” discussion until next week.

For THIS WEEK, take a look at the two YouTube clips that I’ve inserted (see above), and see where YOU think that Sophia is:

• Understanding your comment (doing natural language processing, and mapping what you’re saying against her internal knowledge representation), and coming up with an appropriate non-canned response, or
• Understanding your comment enough so that it triggers a scripted response (e.g., Sophia saying that she wants to “go to college, start a business, etc.”), or
• Understanding your comment enough so she knows that she doesn’t have either a knowledge-based OR canned response, so she gives you a default response. (Remember what that “default” is? See the second YouTube clip.)

Feel free to put your Sophia-assessment in the Comments, below. I’ll do a little read-and-respond. (Sounds a lot like the weekly online discussions that we have in my AI classes at Northwestern!)

So, the purpose of this post is not to do any Sophia-bashing, or even to “game” Sophia and figure out when she’s actually responding in some sort of real give-and-take (even if it’s just discussing the weather), and when she’s giving a canned response.

Our real purpose is to use the current (early-2020) Sophia as a sort of emotional-AI benchmark. A sort of emotional intelligence for AIs and robotic systems.

From where Sophia is today, we can easily imagine next-generation Sophia’s over the next several years.

We can also easily imagine practical applications for empathy-robots.

For those of us who have family members isolated by the COVID-19 virus, we likely wish that we could be there for them. We know that they feel lonely, and feeling lonely leads to feeling scared. Feeling scared stresses their immune system, and that puts them more at risk.

We can stay in touch via phone and various online conferences and chats. Still … that’s not the same as being present, in-person.

Sophia won’t be a replacement for real human connection. Not for a long time, and not to the extent that some recent movies have suggested.

Even so, we can use Sophia as a benchmark. (Or we can define a class of empathetic robots.)

We can foresee a future – not more than a few years away – in which robots not only build things in factories, and do some chores about the house, but also provide some level of emotional support and connection.

This is a new role for robotics, and is – indeed – a brave new world.

What are your thoughts on empathetic robots? Would you like one? Would you like to have the Amazon-of-the-future deliver one to someone in your family?

Live free or die, my friend –

AJ Maren

Live free or die: Death is not the worst of evils.
Attr. to Gen. John Stark, American Revolutionary War