Chatbots — with their artificial intelligence capabilities — were designed to help humans. At least, that’s what many of their developers say they were commissioned to do. But sometimes, maybe they’re too helpful.
Recently, a TV news report in San Diego turned into “too much help.” CW6 ran a story about a six-year-old girl accidentally ordering a $170 dollhouse with the help of Amazon’s Alexa. As the segment wrapped up, anchor Jim Patton was chatting and reflecting on the story by saying: “I love the little girl, saying ‘Alexa ordered me a dollhouse.’”
Turns out, all the Echos and Alexas in San Diego were listening and started ordering dollhouses. Regardless of whether all those Alexas were successful in their purchases, it begs the question if bots are being too helpful and perhaps collecting too much data from people.
“It seems like something out of a 1950s spy movie, where different parts of a home are ‘bugged’ to listen in. Connected devices, such as Alexa and Siri, are listening to wait on particular commands,” said Matt Tumbleson, founder and CEO of Teckst, a platform for customer service teams to perform two-way texting with their customers, including intelligent bots to help customer service agents reply to customers faster. “The scary part about this is that, in order to hear those particular phrases, such as ‘play a song’ or ‘turn on the lights,’ they have to analyze every word spoken to it to determine if it’s a part of a command.”
But Alexa isn’t alone in this listening game. This week, children’s nannies may be fearing a pink slip as Mattel has just unveiled the world’s first smart baby monitor named Aristotle. She — yes “she” — is similar to Alexa and Siri but specifically focused on children.
Josh March, founder and CEO of Conversocial, said nannies have time, and there’s no need to fret.
“In fact, it’s likely to be the opposite — better AI will take more and more white-collar jobs, but they can’t replace the human touch, so there are likely to be more people in caring roles in the future,” said March. “However, these kind of tools will likely make their jobs easier, as well as giving more oversight to parents.”
So, while nannies can relax, there is the concern that bot and AI services may be listening to those interactions, whether users are OK with it or not. In fact, Amazon’s terms of service state that conversations may be recorded and stored outside of the U.S., where privacy laws in another country may be more relaxed. Ultimately, that drives the concern that hackers have potential access to perhaps — who knows — “spy” on those conversations.
But keep calm, experts say there are ways to avoid that potential “spying.” The first is to understand what kind of data is being collected, and the second is to set the devices to certain settings to reduce the amount of data stored on the device or server.
“Many laws around advertising to children require a parent’s consent before marketing to them,” said Tumbleson. “It’s just easier for advertisers and marketers to not target children and instead target their parents. How exactly a company will know if data collected is from a child or an adult requires the company to collect data from both, which could violate many laws and guidance.”
March agreed, saying there’s really nothing to worry about. “There are unlikely to be privacy concerns around the data being collected — unless these tools are the subject of a major hacking incident.”
Experts say it’s often just balancing convenience with functionality and security, because, at the end of the day, bots were indeed invented to help. And, as for their potential in the retail and eCommerce space, according to new reports, it is higher than thought before.
According to Irish research firm Research and Markets, chatbots will redefine human-computer interaction, with Facebook and WhatsApp to lead in the space. The company released its forecasts for 2017 through 2022, saying: “As the interface between humans and computers evolves from an ‘operational’ interface (websites and traditional apps) to an increasingly more ‘conversational’ interface (chatbots, voice interfaces, etc.), expectations about how humans communicate, consume content, use apps and engage in commerce will change dramatically.”
Ultimately, the report goes on, this impending transformation will affect every aspect of marketing and sales operations for every industry vertical.
Experts like John Forrester, chief marketing officer at Inbenta, which specializes in AI and natural language processing, agree that chatbots will indeed benefit customers, especially when it comes to interacting on their own terms.
“When is the last time you tried to contact a company late at night or weekends? You get the dreaded ‘call back during business hours’. Or you are forced to fill out a request form and wait days to get a response,” said Forrester. “However, brands are struggling to staff up and pay for the high cost of providing direct support.”
He said that, as a result, companies without chatbots are causing issues for themselves, such as hiding phone numbers and forcing customers to search through a website only to find “helpful information” to be unhelpful.
“The last year has seen a massive spike in company interest in deploying chatbots — on their websites, mobile apps and social channels,” said Forrester. “The promise of 24/7 access to a chatbot that understands natural language and has access to a massive database of information is tantalizing.”
And as for that report, Forrester said that the adoption of standalone chatbots will happen faster than the report predicts — easily capturing more than 40 percent of the market.
“If you look at the broad adoption of function-built chatbots, servicing a specific use case, along with general purpose support and pre-sales chatbots, you’ll see a massive spike in chat deployments overall, and a majority will be chatbots,” said Forrester.
He isn’t the only one who feels this way about “standalone” chatbots. Sandeep Kumar, managing director at Synechron, feels that the applications for these bots will grow faster than analysts think, especially for industries like financial services and banking.
“Moreover, the technology can be combined with other techniques, from OCR to NLP, to continue to evolve capabilities and connect with other increasingly data-driven and intelligent systems,” said Kumar. “Context is key here, as most of today’s bots lack context.”
So, ultimately, even if chatbots may be “too helpful” in purchasing dollhouses and kicking out the baby’s nanny, they still have a ways to go, and their helpfulness will evolve — hopefully for both business and consumer benefit.