From Bots to Spimes: Emerging Technologies Offer Early Glimpse of Our A.I. Future
Wednesday, August 1, 2012
Web bots, the “internet of things”, machine learning and other converging technological advancements offer an early glimpse of our artificial intelligence future. And marketers need to start paying attention.
When you think of internet bots, if you ever think of them at all, you probably associate them with spam or other nefarious online activities. But bots can be benign, beautiful, beneficial and even social.
Meet Pentametron, for instance. Pentametron is a simple Twitter bot that combs the platform’s data stream searching for random tweets to form sonnets written in iambic pentameter. The results range from silly to downright haunting:
“This usher Mohawk is the business though!
I have an interview tomorrow!:o
Im getting cyber bullied left and right :/
This is a very dark and lonely night.”
Four random, unrelated tweets from four different, distant tweeps. As Gawker points out, “Like most found poetry, Pentametron's sonnets are semi-sensical, often funny, and sometimes profound.”
One of Pentametron’s human creator’s favorite bots is a virtual grammar stickler, @stealthmountain, that responds to all tweets containing “sneak peak” with an automated “I think you mean ‘sneak peek.’” I only wish I had the tech savvy to build a similar bot correcting people on my personal prose pet peeves, but like @stealthmountain’s Twitter bio says, I guess “I live a sad life.”
But beyond being merely artistic or fun, bots can be beneficial. They can remind us to water our plants, and then thank us when we do. They can help alert us about earthquakes (after they happen, of course). They can recommend music based on our tastes. And they can help us anonymously tweet secrets.
On the surface, these bots might not seem particularly impressive from a technological standpoint. They bring together data sources and perform simple automated tasks.
But bots can be social, too. According to MIT’s Technology Review, a team of researchers - skeptical of marketing companies touting the ability to help clients gain Twitter followers and increase interactions - developed bots that were more social than their human programmers [insert your own programmer joke here].
The bots were coded to do social tasks, such as retweeting and conducting virtual introductions between human tweeps in their networks. Compared to the control period where no bots were used, teams participating in the project increased their followers by an average of 43 percent more when bots were helping them socialize. But one team’s follower connection rate increased an astounding 355 percent with its bot’s assistance. This excellent time-lapse data visualization depicts the bots’ connection rates and network impact.
If bots can assemble poems, correct our grammar and even socialize, what does this mean for marketers? We already use chatbots for online customer service tasks. And for better or worse, brand chatbots in our social media networks are the next logical step.
Consider this video from Cornell researchers of two chatterbots conversing. While clearly chatbots have a long, long way to go, observing two chatbots diverge into philosophical, metaphysical and religious debate seems eerily prophetic. Just for fun, try interacting with Cleverbot, the chatbot software featured in the video.
As bots evolve and the “internet of things” inevitably grows in the years ahead, there will be countless opportunities for brands to provide utility in our everyday lives. Chips are becoming more ubiquitous, powerful and affordable, and soon everyday household items - from your alarm clock to your shoes or your car keys - will be networked to the web. Coffee companies, for instance, should note the very straightforward and relatable example Cisco uses in this infographic illustrating how the “internet of things” might work (use your cursor to enlarge and view).
Nike+ is one real-world example of the “internet of things” already in market. FedEx’s Senseaware is another. Google, HP, Cisco and IBM are all advancing technologies based on the concept. And recently CIA Director David Petraeus enthusiastically predicted his agency soon will be able to spy on persons of interest through everyday household objects.
But if you really want to lose some sleep, ponder the confluence of evolving bot technologies, the “internet of things,” and spimes. A spime, currently a theoretical technology but one that’s easy to envision given other existing and emerging technologies, tracks an object from its inception to its obsolescence. The following theoretical example of a spime could be an indication of Nike+’s not-too-distant future:
“A tennis shoe can be thought of as an object in the manufacturing cycle — it first exists as a digital specification for a shoe, then raw materials are gathered and formed into the shoe, an RFID may be embedded into the fabric, and then it is sold. Location and searching for this shoe may involve asking a computer search engine ‘where are my shoes?’ To which the reply may be ‘your shoes are under your bed,’ which would combine identification, location, and data mining (linking the shoes to your ownership of them). Once the soles wear out, the shoes may be sent back to the manufacturer, who will break them down back into raw material which could be used to fabricate a new pair of shoes for you.”
GPS, RFID tags, on-demand manufacturing, cradle-to-cradle product cycles, ubiquitous connectivity, bots, the internet of things and machine learning...One day every day objects will become connected, start to interact with one another, use contextual information to automate tasks for us, and collect data about us and how we use them. Who knows, some people may even opt to become networked themselves. The FDA recently approved the VeriChip, an implantable RFID tag for medical use that could have many other applications.
It may all sound like science fiction today. But increasingly we will be able to track a product throughout its entire lifecycle, learn from its environment along the way, automate personalized interactions with its users, and refine future products and communications based on immense amounts of customer data. As McKinsey puts it, “When objects can both sense the environment and communicate, they become tools for understanding complexity and responding to it swiftly.”
It isn’t a matter of if, but a matter of when. Will you be ready?