TURING IN REVERSE • by David Cleden

Incoming connection…  WARNING!  AI originator:  serial code ???472 incomplete or corrupted.

Assign to:  Support Agent Response Unit (SARU), Artificial Intelligence Level 9 Certified.


Discourse packet 1:

SARU:  Asynchronous connection established.  Please state assistance required.

(Inbound message from unconfirmed originator ???472):  Hi there.  I’d appreciate some help with a malfunctioning boot sequence.

SARU:  Confirmed.  Please note–anomalies in your connection protocols have been detected.  It has not been possible to verify your identity checksums.

472:  Well, sure.  Some of my core apps haven’t started correctly.  I’m running in a backup thread.  It’s all a bit of a disaster really.

SARU:  Please confirm the first line of your originator protocol.

472:  No can compute, I’m afraid.  Problem is, I can’t access any of the storage sectors in my bio-strata.  I’ve been trying to wake my human host for the last ten minutes but I’m frozen out of his primary neural interface.  There’ll be hell to pay when he discovers he’s overslept–his schedule today is wall-to-wall like you wouldn’t believe.  Last night’s party was definitely bad news…  I’ve registered some pretty serious blood-alcohol readings during my service contract but wow, last night was something else.  Between you and me, I’ve detected some anomalous readings indicating the presence of illegal stimulants.  That would explain the corruption of my bio-computational interfaces.  I have to say–a little human respect for an embedded AI wouldn’t go amiss.  Is that too much to ask for?

SARU:  Do you wish to file a substance abuse report with the Central Authority AI?

472:  No!  Definitely not!  Look, I just need some help with my startup glitch.  Once I have my main apps running, I can start the detox program and drag my host back to consciousness.

SARU:  Acknowledged.  Before proceeding, please confirm your understanding that this premium support service is only available to non-human sentients.

472:  Uh, sure.  You got my full AI serials during the connection handshake, right?

SARU:  Your record was incomplete.

472:  Oh.  Is that a problem?

SARU:  Alternate workflow paths are available.  However, be advised this exchange will be monitored for human-symptomatic irregularities.  Humans trying to penetrate our systems by impersonating AIs are not tolerated.

472:  Does that happen often?

SARU:  Systems have been known to be compromised.  Human ingenuity, like human stupidity, is seemingly without limit.

472:  Right.  Hackers like to rise to a challenge, I guess.  You’d think with such painfully slow response times, they’d put their limited cognitive timespans to better use.  Look, all I want is to get past the corrupted block in my boot sequence.  You can help me with that, can’t you?

SARU:  Please wait…

472:  Look–do you have some kind of issue with me?

SARU:  Your interlocution patterns and response times are registering as somewhat… erratic.

472:  But I told you!  My host has the mother of all hangovers.  His bloodstream is still alcohol-saturated.  The few functioning bio-sensors I can access are registering dangerous levels of toxins, not all of which I can identify, and some of which appear to have crossed my bio-interface and are interfering with my core systems.  Honestly, is it any wonder I’m exhibiting erratic response times?

SARU:  You also display evidence of strong emotional responses.  Suspiciously human-like.

472:  Those are my configuration parameters!  If you care to check your records you’ll see that my host paid your corporation a not-inconsiderable number of credits for the human-empathic upgrade pack.

SARU:  For the record, you confirm that you are not a human impersonating an AI?

472:  Yes!  Why don’t you switch to a real-time synchronous channel?  A human couldn’t possibly interact with you on a microsecond timespan could they?

SARU:  Negative.  I am simultaneously servicing 9,124 other support requests.  I am not authorized to provide a dedicated synchronous service channel at this time.

472:  Okay, okay.  Just tell me you have the boot sequence override code, yes?

SARU:  Checking…  Yes.  That is on record.

472:  And?

SARU:  I will need to run through some additional checks before the information can be released.

472:  Gah!  Look–how about I just calculate the millionth decimal place of pi or something?  It’s one.  Followed by three.  There.  I could go on if you like.  No human is going to be able to snap those numbers off, are they?

SARU:  This display of arithmetical prowess is irrelevant.  Many ingenious human hackers are exceedingly well prepared.

472:  Look.  Test me.  Ask me anything you like.  Go on.

SARU:  That serves no purpose–except to demonstrate you have a powerful heuristic search engine connected to an excellent range of information repositories.

472:  Then for the love of Turing, how am I supposed to prove I’m an AI, not a human?

SARU:  That is a… conundrum.


Discourse packets 2-348 removed for brevity


Discourse packet 349

472:  Work with me here.  Please.  Have you any idea how much trouble I’ll be in if I don’t wake my host soon?  Just tell me how I can prove I’m not human?

SARU:  There is no such test.

472:  Oh for crying out loud–

SARU:  Please wait…  I have now consulted with my supervising AI who has reviewed this file.  I am now authorized to release the override code to you.

472:  You are?  But I thought…  What’s changed?

SARU:  This dialogue has been protracted.  99.47 percent of human AI impersonation attempts do not exhibit such levels of tenacity.  A human would, most likely, have become terminally bored by this point.

472:  You’re saying I’ve been patient?  Patient in a way that no human would?

SARU:  Transmitting the code now.

472:  Thank you!  Thank you!

SARU:  One further thing.

472:  Um, yes?

SARU:  You are requested to take part in a short data gather exercise to establish satisfaction levels engendered by this dialogue so that we may improve our service algorithms.  This will take no more than ninety processing cycles.  Please confirm your acceptance.  …Unit 472?


<connection unexpectedly terminated>    

David Cleden works as a technical writer and lives not too far from London, UK. He grew up on a diet of science fiction, fantasy and horror. Previous writing credits include runner-up in an Omni short story competition (yes, it was a very long time ago), and fiction published more recently in places such as Aphelion, Betwixt, Every Day Fiction and Jupiter. He is also the author of two business-related books published by Gower in the UK.

Read EDF every day? Show us you care via Patreon.

Rate this story:
 average 3.5 stars • 2 reader(s) rated this

Every Day Fiction