What Do You Think About AI Reading Emotions in Video Conferences? [Next.Jjin]

[Image source=Yonhap News]

[Asia Economy Reporter Jeong Hyunjin] Have you ever encountered a moment when your boss appeared on your laptop screen, and you wanted to know their emotional state but found it impossible to figure out? When working face-to-face in the office, you could anticipate your boss’s circumstances beforehand and pick up on clues like coughing sounds or behaviors to gauge their emotions. However, during remote or work-from-home situations, nonverbal information is limited to facial expressions and voice tone, making it difficult to understand emotions through the screen alone.

Why Did Zoom Bring Out Emotion AI?

Zoom, a video conferencing service platform that emerged as a major beneficiary during the COVID-19 pandemic, launched a service called Zoom IQ last month to overcome these limitations. This service uses artificial intelligence (AI) to analyze the emotions and engagement levels of participants in video meetings. It detects whether reactions during conversations are positive or negative. Especially when sales teams meet customers via Zoom, it helps analyze subtle reactions and receive feedback.

The service aggregates metrics such as the time it takes for customers to respond and tracks the use of certain words like “um,” “uh,” “good,” or “oh,” which some studies suggest may negatively impact sales success rates. Through this, it analyzes which questions tend to elicit positive responses, facilitating smoother communication and creating an environment favorable for closing deals. This is an ambitious service launched by Zoom, a video conferencing platform, as the COVID-19 era is nearing its end.

Zoom appears to have introduced this service to secure corporate clients and lay the groundwork for another leap forward. In its earnings announcement last March, Zoom emphasized corporate customers and declared its transformation from a simple video conferencing company into a platform company. The goal is to help companies conducting business via video meetings improve their performance with this service. Zoom’s stock price, which was under $100 before COVID-19, soared to $559 in October 2020 but has since dropped to the $90 range, close to pre-pandemic levels. Emotion AI is interpreted as an effort to find future growth engines.

'Over 20 Years of History' Emotion AI Enters Video Conferencing

Emotion analysis through AI is not new. It began in 1995 with the development of AI capable of processing, understanding, and replicating human emotions to enable natural communication between humans and machines. It is already actively used in various fields, including advertising and marketing. However, the COVID-19 pandemic increased the frequency of video meetings and the demand to understand the emotions of participants on screen, accelerating the entry of emotion AI into video conferencing.

Intel is also one of the companies recently showing interest in emotion AI. According to the tech media Protocol, Intel has partnered with Class Technology, a startup developing video class software, to experiment with emotion AI that can assess students’ emotional states. By analyzing students’ facial expressions, they determine whether students are bored or understanding the material, helping to find better teaching methods.

"Concerns Over Bias and Privacy Breaches" Criticism Pours In

The problem is that emotion AI is surrounded by significant controversy. It analyzes emotional states based on facial expressions, voice tone, and word usage frequency, but not all humans express the same emotions with identical expressions. Considering sociocultural understanding and individual diversity, it is difficult to create emotion AI without establishing a “normal” standard, which raises concerns about discrimination and bias. At the same time, there are worries about personal information leaks.

According to Bloomberg, more than 25 civil society organizations, including the American Civil Liberties Union (ACLU) and the Electronic Privacy Information Center (EPIC), recently sent a letter to Zoom CEO Eric Yuan warning that emotion analysis is not scientifically supported, promotes discrimination, and risks privacy exposure. The groups also launched a website to collect signatures from people who agree to block Zoom’s new feature. They do not trust machines to accurately assess human emotional states and are deeply concerned about the side effects.

However, some argue, as reported by U.S. media Axios, that if privacy safeguards are established and the technology is used restrictively, it could be helpful. The technology could assist humans in better understanding and functioning. For example, an automated phone system detecting customer anger or frustration could connect them to a human agent, making it a practical application in daily life.

The controversy is expected to continue for some time. It will be interesting to see whether Zoom and other companies can devise ways to overcome these issues and turn emotion AI into a useful feature for video conferencing.

Editor's Note[Next.Jjin] is a column that conveys news about major companies’ preparations for the future, meaning “looking ahead to the real next of business.” We will uncover and explain both prominent big issues and small but important hidden issues in an easy-to-understand manner.