Is That Your Bot, or Are You Happy to See Me?
Oh, bother. "Only 35 percent of the average Twitter user's followers are real people," suggests a recent article in The New York Times. The rest? Bots or robots, of course. It seems that email bots, spam bots and other algorithmically driven software entities have been adjusted to hold conversations (too strong a word, really) -- to tweet and retweet with us mortals.
In my naiveté I wondered but did not put a lot of mental energy into how my blog could get so many idiomatically correct and totally nonsensical comments such as these:
"Too many folks have already thrown away plenty of good money on nothing but useless salt tablets being shipped from South America. For many working to lose weight, one failure is enough to get them off the path to success. Hence, broccoli is a must-add ingredient in your vegetarian weight loss diet."
"To finish away this fresh lovers is the 1 Cent heel logo and pull tab for that perfect mix involved with past meets produce. The type of Nike air max 96 360 is rebounding next year with some new colorways having a smack of Hyperfuse technology."
Now I know, but somehow, I was expecting more than that -- or I was expecting more from computers, to be honest.
Some days there's more Louis Vuitton on my blog than there is at the America's Cup trials on San Francisco Bay. I secretly thought it must just be the Chinese, which it could be, but I was hoping for real people and not simply computer algorithm-arrhea.
But wait -- there's more (sadly). More than half of Internet traffic comes from nonhuman sources, e.g. Bots, the same Times article says. Finally, bots have come down the cost curve in really impressive style and today, "mercenary armies of bots can be bought on the Web for as little as US$250," which can only lend greater credence to the T-shirt aphorism, "Beam me up, Scotty. There's no intelligent life here."
Of course, this is an arms race, and already there are vendors building anti-bot bots that will engage bots in pseudo-conversations, eventually distracting them into their own chat room/death spiral. That's somewhat comforting, but not much.
I remember being quite impressed, at first, with the possibilities of bots handling routine matters like customer triage when people log into a service site, for instance. However, the low numbers of real people involved in Internet activities already scares me. This is a sort of Internet of Things that no one ever planned on.
If only 35 percent of your followers are real, and half of the traffic on the Internet is fake, and about 82 percent of the email is either spam or bacn, then this seriously diminishes the Net as a business tool and calls into question the logic of moving everything online.
Bots Gone Wild
OK, seriously, moving online is still the right move for all kinds of reasons, but the amount of additional friction we're beginning to encounter provides significant headwind to grand Internet strategies and will most likely add cost to what we do there, not to mention the rising risk of fraud.
Of course, this also means an opportunity for entrepreneurial developers who scheme to produce another layer of security software to handle this problem. Come to think of it, with numbers like these, I wonder if the NSA's spy program is being largely consumed by bots and not by bad guys calling home to some place in a desert. True, the NSA is dealing with phones, but suppose there's a separate program that monitors Internet traffic?
Will the sheer volume of chatter on the Internet render the monitoring less effective? What if there is no program to monitor the Internet? There should be one, and it would be more useful to get the bad guys out of the bot game than chasing innocent business calls. You know?
It's only a matter of time before some bad guys build bots that impersonate a Coke machine calling home for more syrup or (gasp!) a jet engine reporting a problem with a bearing. As the Internet of Things comes into view, the obstacles and challenges become more clear. I hope someone is looking into this.