Lemonade, an AI-powered insurance company received a backlash on Twitter around what it claims to be “a poorly worded tweet” from its Twitter thread. The thread explained how Lemonde makes use of bots and machine learning algorithms to create a seamless, instant and delightful experience for its customer
Lemonade also shared how their data collection mechanism puts them at an advantage versus traditional insurers. They claim to make use of the data to help them understand the level of risk each customer brings, which improves their underwriting, customer acquisition, and fraud detection capabilities.
“For example, when a user files a claim, they record a video on their phone and explain what happend”
“Our AI carefully analyses these videos for signs of fraud. it can pick up non-verbal cues that traditional insures can’t, since they don’t use a digital claims process”
“This ultimately helps us lower our loss ratio (aka, how much we pay out in claims vs. how much we take in), and our overall operating costs.”
“In Q1 2017, our loss ratio was 368%” (friggin’ terrible), and in Q1 2021 it stood at 71%!”
– said @Lemonade_Inc on Twitter
This thread was not received well by the Twitter community with the bulk of the conflict centered around Lemonde’s AI picking up on non-verbal cues to detect signs of fraud. One follower said in a quoted tweet
“Imagine that you’re an autistic person whose home just burned down. You’ve just gone through the most stressful event of your life—and now you have to worry about policing your own speech patterns so your insurance company doesn’t flag your claim as fraud,”
There is also the fact that facial recognition systems like the one touted out by Lemonade have—by and large—failed people of color. There’s a crushing amount of evidence detailing how the facial recognition tech that’s used in housing applications, airport security, and the entire criminal justice system inadvertently misclassifies non-white faces. Naturally, Lemonade’s followers were sure to mention that as well.
Lemonade has since deleted the series of tweets
“We deleted this awful thread which caused more confusion than anything else,” the company said in a blog post Wednesday, adding that using the phrase “non-verbal cues” was a poor choice of words.
The company said it doesn’t use technology that relies on disproven pseudoscience like phrenology and physiognomy (theories that purport a link between physical appearance and intelligence or character), nor does it evaluate insurance claims based on customer “background, gender, appearance, skin tone, disability, or any physical characteristic.”
Lemonade clarified their AI uses a facial recognition technology to highlight claims submitted by “the same person under different identities,” such as a person wearing disguises to submit multiple claims, adding that flagged claims are later reviewed by humans and no claim is ever auto-denied based on AI.
Lemonade launched in 2015 and was seen as a disruptor in the insurance business, being paper-free and using artificial intelligence and chatbots to help process insurance claims. Lemonade asks customers submitting claims to include a video of them explaining their situation in their own words, partly because the company believes it could reduce fraud. “Behavioral economics research … has shown that we humans are less prone to lying when we’re looking at ourselves speaking in a mirror [or] selfie camera,” the company said in a blog post-Wednesday. The insurance firm operates in the U.S., Germany, the Netherlands and France, with plans to expand. In May, the company’s first-quarter earnings report revealed Lemonade saw a net loss of $49 million, partially driven by insurance claims after the record-breaking storms in Texas earlier this year. Lemonade expects to earn between $117 million and $120 million through the rest of the year. Last month, the company announced plans to enter the auto insurance market.