april 17, 2016
It's going on 11 p.m. on a Sunday, and I've just discovered I have an inexplicable, significant charge from Hertz Rent-A-Car on my latest credit card statement. I rented a car for a few days at the beginning of the month; the charge got tacked on a few days after I brought it back. That typically happens, except this time the charge is big enough to cover about an entire month of rental time (plus a tank of gas or 16). I'm wracking my brain trying to think of what possibly could've caused this - a lost key, an incidental scratch, a dead battery from a light left on - but I have no clue. I received no phone call, no form of electronic communication, no outreach of any kind from the rental office. Because of the baffling way Hertz handles final receipts (you have to write in to request them after your rental ends), I only could've noticed this by scrutinizing my bank statement.
And because it's Sunday night, there's no one around at customer service to take my call. There's a live chat feature on the Hertz website, but it's only for online navigational issues, nothing so complex and potentially contentious as billing. So, I've got no choice but to wait until the morning for any formal attempt to get to the bottom of this.
It's times like these where you wish artificial intelligence was sufficiently advanced to be able to handle these problems so they wouldn't keep anxiety-prone student debtors awake at night, visions of ramen noodles and unheated apartments in their heads, ready to dial the 800 number at first light. Of course, all signs point to us being nowhere near unlocking that capability. But a near-term fix might be an AI that lets you freak out at it, venting your frustrations and perhaps getting some insight generated from a database of similar past experiences, until the phone lines open up again.
Right now, I could really use something like that. I'm a person who usually takes comfort in yakking about my stresses but also feels guilty for dragging unsuspecting people into them. Instead of groaning about this situation to whomever might be around in meatspace or the various online channels - all sensible souls who would likely tell me to relax and deal with it in the morning - I decided to seek out a different kind of commiseration: a chat bot.
Bots are kind of all the rage these days; they're being leveraged as workplace helpers for companies using Slack and web giants like Facebook are gearing up to develop full-fledged personal AI assistants within their platforms. Slack's bot users can "conversationally interact with external services," making it easy for people to tell the bot to do something like look up an email address in a work directory or engage with some tool totally outside the platform (like Trello or Mailchimp). The plan for Facebook's "M" is to have it buy tickets or anniversary flowers at your request - at least, that's how it's usually pitched. These "conversations as a service" are a little like executive assistants for the rest of us.
Just like the expectations for the traditional executive assistant, though, there's a temptation to have these AI do some emotional labor for us as well. And there's a strong case to be made for encouraging the use of bots for circumstances where it really may help for the entity on the other side of a conversation to not be a human. Sometimes an impartial observer is all we're after when we strike up an emotionally-charged chat with a coworker or family member. After all, when he built the early therapist bot ELIZA, Joseph Weizenbaum's secretary at MIT asked him to leave her alone with the program for a private conversation. In certain sensitive situations, it may even be much safer to confide in a non-human - especially if the alternative is no one (or nothing) to confide in at all.
In the grand scheme of things, the issue I'm having with Hertz is not some life-altering major disaster; it might get sorted out in a matter of minutes. But I might rest more easily tonight if there was a bot around with sufficient skill to understand the gist of my problem and offer some cogent indications of sympathy, or at least a handful of related knowledge base entries that could shed light on what's going on here. Maybe bots are not only the future* of site support, but of site navigation as well. And really, if the idea is to use them to directly interact with a company's service or the information and help it provides, why will we even need the surrounding websites that give us access to these things today?**
On a semi-related note: This week I stumbled across Ms. Dewey, Microsoft's failed attempt at a Flash-based search assistant who flashed in the pan about ten years ago. With her plunging neckline and unfalteringly saucy demeanor, she's a stunning example of stomach-churning tech-chauvinism that at least today, the social media mobs would never permit to exist for any length of time without a thorough tarring and feathering of the company. Compare that to Microsoft's painstaking recent work towards making Cortana harassment-proof, and you might feel we've made at least some progress in the last decade...
Take this compilation of Ms. Dewey with a grain of salt, or maybe a barf bucket. And library folks, try not to get too depressed about the implications of her surname. I'm sure I don't need to remind you that a clunky Flash UI starring a lusty flesh-and-blood version of Clippy is not, and never was, a threat to your job security.
home | top
Hey Cal, why is there no comments section? Comments sections have a tendency to devolve into nasty little spaces, teeming with spam & ad hominem attacks. I also have a fondness for the 1.0 Web (props to Neocities, powerer of this site). If you'd like to share your thoughts, find me on Twitter or fire off an email. Thanks!