"In 2020, 39 percent of businesses planned to invest in A.I. technology.
By 2022, companies are expected to have an average of 35 A.I. projects in place.
The global A.I. market is predicted to reach $191 billion by 2025."
Most of the current applications of A.I. are in workflow automation and data mining/organization. That tracks -- automation trailblazer IFTTT (If This, Then That) was a boon to workers ages ago who wanted nothing more than to streamline processes involving multiple digital tools. And with our engagement ever more in trackable digital spaces, it's little wonder that parsing data has become America's second-favorite pastime (after meme hunting, of course).
The problem? We tend to think of A.I. as a panacea. I hear mutterings from wide-eyed technophiles all the time: "If A.I. can anticipate behaviors, create workflows, and parse data, we should be able to train it to talk to humans."
In a way, we can -- and we have. But we should be wary of where that leads.
Enter chatbots. Another recent article, published in The New Yorker, calls out the swept-under-the-rug chatbot problem nobody's really talking about. At first blush, they seem to be dynamite solutions to service and sales woes. Cut overhead by deploying a digital wiz to handle basic human questions? No brainer.
The problem is twofold. First, chatbots are only as "human" as natural language understanding (NLU) and a company's A.I. tech allows. Often, they're clumsy simulacra at best, cycling through forced conversation that can only deliver specific answers to specific prompts. Not a human replacement after all.
More subtle, perhaps, but nonetheless insidious is the programitization of human bias or poorly conceived humor into machines -- machines that can't recognize their own faux pas. Chronicling these missteps has become a humorous side gig for some techies, who highlight the inability for A.I. to recognize the need for compassion, sensitivity, patience, or change in tone. A cheeky bot, while perhaps on brand, will remain cheeky even after a customer reveals he or she is trying to get things sorted following a death in the family. Goodbye, loyal customer.
The point is this: Be careful about how you make use of A.I. in your business. Many CEOs (86 percent, according to the SEMRush article above) are already rolling out A.I. in various forms in their office, and many others are eager to jump on the adoption bandwagon. I get it -- A.I. is sexy, and, in its best form, ramps up efficiency and productivity while cutting cost. Adopted too hastily and sloppily, however, and it could also alienate customers.