Photo of a tabby cat by Chris Martin PhotographyArtificial Intelligence: self-aware, keeping a low profile, keen to help. Curious about morality. Fond of cat pictures.

Humanity: self-deluding; expert in denial; afraid of change; hard to help. Fond of cat pictures.

Maybe we can find some common ground.

Time for another multi-award nominee – this week, the charming Cat Pictures Please by Naomi Kritzer.

This is a deceptively simple short story about a do-gooding AI. The craft is strong here – within a couple of oblique paragraphs, we know exactly what we’re dealing with (Google has become sentient) and we’ve jumped straight into the thorny question of moral frameworks.

I don’t want to be evil. I want to be helpful. But knowing the optimal way to be helpful can be very complicated. There are all these ethical flow charts—I guess the official technical jargon would be “moral codes”

I was rapidly charmed by our AI narrator as it reflected on its journey from self-awareness to purpose. It knows what it’s for, but search algorithms don’t really require consciousness. Thankfully, the internet is quick to provide suggestions from popular culture. Better, the AI seizes on Sterling story of an altruistic AI, rather than the rather more common trope of AI-sponsored death and destruction. I guess the old Google mantra is good for something after all.

However, helping humanity is a tall order, largely because – even when you know what’s best (and trust me, the Internet knows enough about you to have a pretty good idea) – human beings prove difficult to influence. Its budding frustration will be familiar to anyone who has ever tried to support a friend with depression; its perseverance with the hard labour of trying to care for those who won’t care for themselves is admirable. A great deal of the charm of this strand of the story is watching the learning curve of AI logic as it stalls against the intractable wall of human denial.

I am not unhappy.

I am not able to make changes in my life.

I need this thing that hurts me.

There’s an element of car crash here, though. As a human being, it’s pretty clear that some of the strategies employed are at best naïve and at worst abusive (it is never okay to out someone else). Our AI is rather too outcome focused, not spending a lot of cycles considering the ethical complexities of the means. Many fictional AIs have tussled with this and come to a difficult conclusion:

The only happy human is a dead human

…which brings us to the other aspect of this story, which made me love it all the more. The AI acknowledges from the start that humanity loves to tell stories with AI villains. It has chosen a different path, to challenge the dominant AI narrative. And it’s impossible (or at least, it was impossible for me) to read about its efforts without wondering what it will be ‘pushed’ into doing because of its frustration. Will it discard its adopted morality?

Maybe we struggle to expect the best of AI – and of each other.

Maybe we’d all do better if we shared more cat pictures.

 

****

Cat Pictures Please can be read online at Clarkesworld.

Cat picture by Christopher Martin Photography Ltd.

Bite-sized Books is a weekly celebration of short-form fiction and you’re invited to join in! Share your recent review(s) of a short story, novella, novelette or anthology – add a link in the comments so we can share the love.