Eye on AI - May 28th, 2021
Welcome to Aigora's "Eye on AI" series, where we round up exciting news at the intersection of consumer science and artificial intelligence!
This week, we’ll be discussing the release of LimeChat’s first-of-its-kind Level 3 chatbot that uses big data to understand intent, then transition to a new AI that could help revive the struggling airline industry.
First-of-its-Kind Level 3 Chatbot ID’s Language, Sentiment & Intent
We begin with some encouraging news about new developments in the retail chatbot. According to the Yourstory article “Using AI and chatbots, this startup is helping D2C brands build a compelling buying experience”, AI startup LimeChat has developed the first-of-its-kind Level 3 chatbot after consulting with hundreds of global direct-to-consumer (D2C) brands.
“Most chatbots can only take in fixed predefined responses and are unable to answer questions that have not already been programmed,” writes Yourstory contributor Trisha Medhi. “... LimeChat bot can take in free-flowing text answers, one-word responses, spelling mistakes, etc., and still understand what a customer is trying to convey to provide the most relevant information.”
Think of it this way: where previous versions of D2C chatbots had relied on pre-programmed questions that were often misinterpreted due to spelling mistakes, oddly-phrased questioning, etc., LimeChat’s system was developed to understand user intent despite questioning missteps; where previous chatbot versions asked users to re-phrase questions that were misunderstood, LimeChat provides simple multiple-choice questions for clarification regarding the specific product or service in question.
“LimeChat bot can take in free-flowing text answers, one-word responses, spelling mistakes, etc., and still understand what a customer is trying to convey to provide the most relevant information,” writes Yourstory contributor Trisha Medhi. “...LimeChat’s Level 3 bot can give a 53 percent higher engagement rate than a Level 2 bot.”
As the chatbot market continues to grow,––Grand View Research reported the global chatbot market is projected to reach $123 billion by 2025––LimeChat’s announcement could mean big news for its market share.
It should be noted that LimeChat’s claim to have created the first-of-its-kind level 3 chatbot was made by the company’s founders, not independent researchers. Yet even with other level 3 chatbots likely to hit the market soon (if they’re not already available), this news could lead to an accelerated customer base for LimeChat.
New AI Helps Plan Flights and Ease Stress in Air Traffic Controllers
Let’s continue by lifting our focus toward the skies. According to the Time article “Travel Is Coming Back, and Artificial Intelligence May Be Planning Your Next Flight”, Flyway, a new flight planning AI, is currently being tested by Alaska Airlines to assist air traffic controllers in directing flight routes, helping to ease the pressure off what was once largely considered to be the most stressful job in the world.
“[Flyway is] the product of a Silicon Valley company called Airspace Intelligence, which is hoping to make air travel more efficient and safer as airlines try to recover from a year in which the pandemic ravaged the travel industry,” writes Time contributor Alana Semuels. “The idea of AI setting paths for jets streaking through the skies at hundreds of miles per hour might sound terrifying, but Flyways does not dictate—it advises, and humans always have the final say.”
Flyway was developed by Phillip Buckendorf, the 30-year-old CEO and co-founder of Airspace Intelligence, when he saw much pressure dispatchers were under to deliver zero-fail results using rudimentary technology. Dispatchers mostly rely on paper printouts to see weather forecasts. They determine things like wind speeds, storm probabilities, etc. by decoding strings of letters and numbers on a federal website. It’s a tedious, time-consuming, and extremely high-pressure job that relies on data and prediction, two things AI is known for.
Alaska Airlines eventually agreed to test Buckendorf’s Flyway for a six-month trial in 2020 on all of its continental flights. During that time, Flyway reduced flight times and fuel on 64% of the flights, cut an average of 5.3 minutes off each Alaska Airlines flight, and saved a total of 480,000 gallons of fuel, avoiding 4,600 tons of carbon emissions in the process.
This type of AI could truly be game-changing for the flight industry, which has struggled mightily during the pandemic. Furthermore (this part we love), rather than acting as a replacement for human workers, Flyway was designed to be an assistant. It makes recommendations, that’s all. It’s up to the controller to decide whether or not to act on them.
“‘A man/machine collaboration will do better than just a man or just a machine,” says Pedro Domingos, a professor of computer science at the University of Washington. Deep Blue, an IBM program, may have beaten Gary Kasparov at chess, he says, but ‘centaur chess teams,’ in which humans worked with AI, beat both humans and machines. ‘It’s not that people will be out of jobs, it’s that they will use computers to do their jobs more.’”
Flyway is off to a strong start but is still a ways from being implemented at scale. More testing is needed for something as important as directing airline flights––as we’ve seen before, some AI algorithms are unintentionally infused with bias, which can be especially dangerous for zero-fail applications. But if more applications of AI are adopted by airlines, we could begin to see a revival of this struggling industry.
That's it for now. If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, click on the button below to join our email list. Thanks for stopping by!