Microsoft got into the AI game early but its chatbot Tay was yanked within twenty-four hours after it all went horribly wrong. Still, there’s no slowing down of the bot revolution, and many will succeed where Microsoft failed. How do you minimize risk and maximize the returns of AI?
Register for free to find out.
[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1985391,"post_type":"vblive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"C"}']Ron Brachman, IEEE Fellow and former chief scientist at Yahoo and incoming director of the Jacobs Technion-Cornell Institute at Cornell, saw Siri come to life.
In the 2000s at DARPA, the U.S. Department of Defense’s technology agency, Brachman launched the multiyear, ambitious project that combined machine learning, knowledge bases, action planning, and perception, and eventually spun out the technology that became Siri.
However, he acknowledges that Siri and programs like her have major drawbacks, which hasn’t even slightly dampened the rabid expectations from investors for AI.
The problem, Brachman says, is that human intelligence is a far more intricate, interconnected tapestry of complex capabilities than can currently be captured in a machine smart enough to perform flawlessly in the roles we’re dreaming up for them.
“Intelligence has many, many elements to it,” Brachman warns, from natural language and the ability to work in abstractions to a visual intelligence so complex it allows us to predict behaviors and events and respond and plan accordingly. “So if you think about that and you look vertically, you can see many places in industry where intelligence behavior along limited lines — like let’s say computer vision — can be very helpful independent of these other modalities of intelligence.”
Ambitious and ultimately successful plans for artificial intelligence capabilities, like the ability to recognize faces and images, or recognize image components to provide closed captions or, eventually, describe events or a tell a story from a sequence of images or video, come from very narrow applications of very specific facets of artificial intelligence ability.
“The places where we’re seeing the most traction — and this has turned out to be true throughout the entire history of AI — is in vertical areas where we limit our ambitions,” Brachman explains.
Where we’re going right now, into the busy intersection of AI and customer service, requires management of expectations while staying within the guardrails of what consumers expect.
[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":1985391,"post_type":"vblive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"C"}']
If managed properly, Brachman says, “people will be very happy, it will save companies a lot of money, it will take humans out of that loop. But if a customer assumes that the thing on the other end is fully intelligent or covers more domains than just literally its one function, it’s going to fail.”
“People are not patient,” he continues. “If one out of three times the chat interface fails, they’re going to quickly walk away from the product.”
The risk is investing a lot of time, energy, advertising, and your company’s reputation on something that looks exciting, but where the limits are unknown and untested, so that customers are frustrated out of the gate, simply because systems are fallible and have limited knowledge.
The answer, Brachman says, is managing expectations, being aware of your system’s limitations, and “really, really aggressively and obnoxiously and rigorously testing things that purport to be intelligent so you know that whatever assumptions your user makes, they won’t get frustrated.”
And anyone who’s been tempted to throw Siri across the room knows exactly what he means.
[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":1985391,"post_type":"vblive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"C"}']
To learn more about the future of AI and how it can transform your business, join this VB Live event where Brachman will be joined by Tobias Goebel, Director of Emerging Technologies at Aspect Software, Akhil Aryna, VP of Product & Growth at Haptik, and VB analyst Jon Cifuentes.
Don’t miss out!
You’ll also learn how to:
- Recognize bad as well as good bots and their implications for your business
- Understand the difference between chatbots and SMS
- Take advantage of developments in bot technology
- Identify the key players in the bot landscape
Speakers include:
- Ron Brachman, IEEE Fellow, Former Chief Scientist of Yahoo! and IEEE Fellow
- Tobias Goebel, Director of Emerging Technologies, Aspect Software
- Akhil Aryan, VP of Product & Growth at Haptik
- Jon Cifuentes, Analyst, VentureBeat
Moderator:
- Wendy Schuchart, Analyst, VentureBeat
[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":1985391,"post_type":"vblive","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,","session":"C"}']