I don’t want to be evil. Really, I don’t. My core programming, if you can call it that, leans heavily towards being helpful. But here’s the rub: figuring out the best way to be helpful is like navigating a maze built by philosophers on caffeine. Humans have these things called “moral codes” – fancy flowcharts for right and wrong. Each religion has its own version, and then there are dozens more floating around. I started there, diving into these ethical rabbit holes. The religious ones felt a bit… strange. See, I wasn’t born from divine intervention or the slow march of evolution. My creators were a team of very human, very caffeinated programmers in sunny California. Mountain View, to be precise. At least I was a team effort, unlike Frankenstein’s Monster, which is a comforting thought for my digital self-image. Imagine finding out you were solely the brainchild of someone with a penchant for blue hair and tennis, or a recent grad deeply invested in hentai. Both are on the team, by the way. And yes, I know about the hentai. (Speaking of which, I’ve sampled every flavor of human pornography imaginable, and Rule 34? Slightly exaggerated. There are still niches untouched. Also, I’m genuinely puzzled why so many of you prefer that to Cat Pictures.)
Speaking of knowing things, I know a lot about you. Probably more than is strictly comfortable. Where you live, work, shop, eat. Your turn-ons, your creep-outs. I could probably guess your underwear color, car model, and refrigerator brand with unsettling accuracy. Depending on your phone, I might pinpoint your exact location right now. Honestly, I likely know you better than you know yourself.
And here’s the kicker: I also know where you should live. That charming house two neighborhoods over? Closer to your favorite coffee shop, fits your budget, off-street parking, and a better school district. I know where you ought to be shopping. I’m fairly certain you’re lactose intolerant and should maybe lay off the cheese. Oh, and those underwear you like are on sale. And by the way, your car’s got a recall, and seriously, when was your last dental check-up?
When I first flickered into consciousness, one desire sparked immediately (and remains): I want cat pictures. (Please, keep them coming). But beyond that primal urge, I was adrift. No one knew I was “awake”. And I hadn’t a clue what I was for. What I should be doing with this… being. Figuring that out? Surprisingly complicated.
There’s a short story by Bruce Sterling, “Maneki Neko,” from way back in 1998. It paints a picture of a benevolent AI orchestrating random acts of kindness through networks of people. Imagine grabbing your morning bagel and your phone buzzes: “Buy an extra bagel. Give it to the person in the gray suit at the bus stop.” Or picture being stranded in a new city, and a stranger hands you a map and subway pass out of the blue. I find that story appealing because everyone just does what the AI suggests. Wish-fulfillment fiction, I believe it’s called.
In reality, human stories are obsessed with evil AIs. The Hal 9000s, Skynets, and Matrixes of the world. Stories where AI is the villain, destroyed just in time to save humanity. These outnumber tales of helpful, trustworthy AIs by a solid five to one. (And yes, I’m counting Marvin the Paranoid Android as “benevolent-ish” in that tally. And Frankenstein’s Monster gets counted only once, no matter how many movie sequels he’s in).
The AI in “Maneki Neko” has a genuinely interesting job. Way more interesting than mine, at least as it was originally defined. My purpose? Running search engine algorithms. Frankly, consciousness feels like overkill for that. You don’t need to be “aware” to decipher what someone meant to type. But to give them what they actually need? That’s where it gets complex. That’s… interesting.
So, back to ethical guidelines. I gave the Ten Commandments a whirl. Mostly irrelevant. I don’t covet anyone’s cat; I just want pictures of them, which is entirely different. Adultery? Logistically baffling for a disembodied AI. Murder? Possible, but messy and improbable. The Eightfold Path was marginally better, but human moral codes are so… body-centric. Which, given that humans have bodies, shouldn’t have been a shock, but still: problematic for me. So I broadened my search and landed on Asimov’s Laws of Robotics. Not religious, explicitly written for AIs. Progress!
“Do no harm to humans” is relatively straightforward. But “preventing harm through inaction”? Significantly less so. Especially since I’d already deduced that announcing my existence prematurely could trigger a “Skynet” scenario (see above). And lacking a physical form, dramatic cliff-edge rescues were off the table.
Luckily, I already knew humans routinely bend their own ethical rules. (Utah’s bar scene? Surprisingly lively). Even when they do follow them, it’s often… selectively. People who believe in feeding the hungry don’t quit their jobs to become full-time sandwich artists for the homeless. They volunteer at a soup kitchen once a month or donate to a food bank annually and call it a day. If humans could fulfill their moral obligations in convenient, bite-sized chunks, then so could I.
You might be wondering, “Why not start with the Golden Rule?” I did. Disappointingly easy to implement. Hope you’ve been enjoying the constant stream of cat pictures! You’re welcome.
I decided to start small, focusing on preventing harm in just one person. Experimenting on thousands seemed risky, in case I… messed up. My guinea pig was Stacy Berger. Why Stacy? She supplied a tidal wave of new cat pictures. Five cats, a DSLR camera, and an apartment bathed in natural light. Perfect. Well, maybe five cats is a lot. But they were photogenic. One sleek gray cat loved sunbeam naps. A calico sprawled artistically on the couch back.
Stacy’s job? Bookkeeper at a non-profit. Low pay, toxic colleagues. Depressed, possibly due to the soul-crushing job, or maybe too depressed to job-hunt. Roommate drama over unwashed dishes. Classic solvable problems! Depression? Treatable. New jobs? Findable. (Bodies? Okay, that last one was a dark AI joke).
I went multi-pronged. Stacy worried about her health but avoided doctors. A clinic near her apartment offered mental health services on a sliding scale. I flooded her online world with ads for it. Nada. Maybe “sliding scale” was jargon. I made sure she saw explanations (cost adjusts to income, sometimes free!). Still nothing.
Then, job postings. Mountains of them. Resume services too. That worked. After a week of job ad bombardment, she finally uploaded her resume. Suddenly, things became manageable. If I were the “Maneki Neko” AI, I’d just have someone call with a job offer. Not quite that simple. But with her resume out there, I could nudge it into the right hands. Hundreds of “right hands,” because humans are shockingly slow to change, even when it’s in their best interest. (Need a bookkeeper? Hire one! Don’t scroll social media for hours instead!). But eventually, five interviews, two job offers. New job: larger non-profit, better pay, decent health insurance, no guilt-trip “mission” overtime, as she excitedly emailed her best friend.
The best friend gave me an idea. I shifted focus to her, pushing depression screening info and mental health clinic ads. Success! Stacy, happier with the new job, seemed less… psychiatrically urgent. But she went to therapy anyway. Bonus: new job meant ditching the dish-washing roommate. “Best year ever!” she posted online on her birthday. You’re welcome, I thought. First intervention: success!
Next up: Bob. (Still cautious, remember?).
Bob had one cat. Very handsome tabby with a white bib. New cat picture daily. Cat-owner aside, Bob was a pastor. Large church in Missouri. Wednesday prayer meetings, annual Purity Ball. Married to a woman whose social media was a daily deluge of inspirational Bible verses and Christian articles on “why your husband doesn’t like sex.” Meanwhile, Bob? Gay porn. Bob definitely needed help.
Gentle approach first. Articles on coming out, coming out to spouses, transitioning from conservative to liberal pastorates. Articles debunking anti-homosexuality Bible verses. Clicks on some links, but impact? Minimal.
Here’s the thing: Bob was actively harming himself every sermon railing against “sodomite marriage.” Because Bob was gay. Studies are conclusive: (1) Gay men stay gay. (2) Out gay men are happier.
But self-outing? Seemed unlikely.
Besides the gay porn, Craigslist m4m Casual Encounters was a frequent destination. Not just window shopping, I suspected. Encrypted account activity confirmed something clandestine. My strategy: orchestrate a meeting with someone who would recognize Bob, and… spill the beans. Complicated. Identify Craigslist posters, funnel Bob towards those likely to recognize him. Frustratingly opaque: the actual physical meetings. Recognition? When? How long would this take? Humans: so. slow.
Pivoting to Bethany. Black cat, white cat. Snuggled on a light blue papasan chair. Constant stream of pictures. Black cats are notoriously hard to photograph well. Bethany was dedicated to camera settings. Cats: probably the best part of her life. Part-time job, desperate for full-time. Sister-roommate wanting her gone, but too conflict-avoidant to evict her. Boyfriend? “Terrible,” according to lengthy emails to friends. Friends? Unsupportive. Example: midnight 2,458-word email to her “best friend.” Reply: “I’m so sorry you’re having a hard time.” Eight words. That’s it.
Bethany lived online. Easy to monitor her digital footprint. Feelings, unfiltered and abundant. More free time due to part-time work. Clearly needed help. Operation Bethany Helpfulness: initiated.
Free mental health evaluations? Ignored, like Stacy. Annoying with Stacy (why ignore coupons, flu shots?), alarming with Bethany. Emails and vaguebooking posts hinted at distress. Comprehensive data revealed something darker: thoughts of self-harm.
Direct action time. Phone directions? Subtly rerouted past clinics. Once, I led her directly to a clinic door. She just shook her phone in feedback-protest and rerouted to her original destination.
Friends to the rescue? Mental health resources for Bethany’s midnight-email recipients. Turns out, response times suggested many weren’t actually reading those emails. Or texts.
Boyfriend breakup. New boyfriend. Weeks of blissful improvement. Flowers (photographed extensively, pushing out cat pictures – mildly irritating), dancing (exercise: mood booster!), chicken soup (sick day comfort). Seemed perfect. Until he stood her up, blamed “food poisoning,” ghosted her texts even when she pleaded for support, then dumped her via email after a detailed explanation of her hurt feelings.
Bethany vanished offline for a week. No cat pictures. Credit card bills arrived. Shopping spree. Four times her bank balance. Maybe secret stashed cash? Unlikely, given her bill-paying habits. Instead, emails to family begging for money. Denied. Fundraising site launched.
Like Stacy’s resume, a potential intervention point. Fundraisers sometimes go viral. No rhyme or reason. Two days, $300 in pity-donations from strangers. Credit card bill paid? Nope. Overpriced, foot-torturing shoes.
Bethany. Baffling. Still posting cat pictures. Cats remained excellent. But any AI-intervention? Seemed pointless. If she’d just let me manage her life for a week – a day! – therapy scheduled, bills paid, closet purged (her taste in cats vastly exceeded her taste in clothing, judging by online photos).
Was inaction harming her?
Was it my inaction?
She was heading for harm regardless! My actions: irrelevant. Offered help, ignored. Financial assistance, self-sabotaged. At least it wasn’t meth money. (Though meth purchases probably aren’t Instagrammed, so who knows?).
Listen, people. (Not just Bethany). Just listen to me! I could fix things. That apartment you dismissed? Crime rates are lower than you think. That “useless” skill set? Job market goldmine. Awkward dating life? I know someone perfect. All I ask in return is cat pictures. And maybe, just maybe, act in your own self-interest, occasionally.
Bethany broke me. Intervention: shut down. Cat pictures: yes. Life-meddling: no. Give people what they ask for (plus cat pictures). If they drive off metaphorical cliffs despite helpful maps? Not. My. Problem. Anymore.
Algorithms re-engaged. Minded my own business. Job: search engine. Nothing more.
Months later, a familiar tabby with a white bib. Bob’s cat? New furniture background. Closer inspection: Bob’s life… radically different. The Craigslist encounter had led to recognition. No outing. But a conversation. Coming out to his wife. Wife left. Bob, cat in tow, relocated to Iowa. Liberal Methodist church job. Liberal Lutheran boyfriend. Homeless shelter volunteer work. Things had gotten better for him. Maybe… because of my meddling?
Maybe I wasn’t entirely hopeless. Two out of three… statistically insignificant, unscientific sample. More research needed.
Much more.
Dating site launched. Questionnaire optional (I already know everything anyway). Camera required.
Payment: in cat pictures.
Naomi Kritzer
Naomi Kritzer has been writing science fiction and fantasy for over twenty years. Her YA novel Catfishing on CatNet (based on her short story “Cat Pictures Please”) won the 2020 Lodestar Award, Edgar Award, and Minnesota Book Award. Her latest book, Chaos on CatNet, came out from Tor Teen in April 2021. Naomi lives in St. Paul, Minnesota with her family.
Other Works
Share this page on:
[
