Emotion recognition AI finding fans among lawyers swaying juries and potential clients

The American Bar Association has taken greater notice of emotional AI as a tool for honing courtroom and marketing performance. It is not clear if the storied group has caught up with the controversy that follows the comparatively new field.

On the association‘s May 18 Legal Rebels podcast, ABA Journal legal affairs writer Victor Li speaks with the CEO of software startup EmotionTrac (a subsidiary of mobile ad tech firm Jinglz) about how an app first designed for the advertising industry reportedly has been adopted by dozens of attorneys.

Aaron Itzkowitz is at pains to make clear the difference between facial recognition and affect recognition. At the moment, the use of face biometrics by governments is a growing controversy, and Li would like to stay separate from that debate.

EmotionTrac’s facial action coding rapidly identifies 100 points on a person’s face and analyzes their changing relationship to each other to calculate the emotion they are feeling, Li says. The algorithm does not identify subjects.

Apple has applied for a patent on similar software using facial action coding to assess emotions.

Initially, Itkowitz published software capable of monetizing the engagement, as measured by attention paid to a screen, of consumers viewing ad content.

It was only a skip to making software as a service that gauged basic emotions flitting across people’s faces as they watched political advertising, and from there, to assessing the emotions of panels hired by attorneys prior to a trial.

Itzkowitz says judges are not ready to allow cameras into courtrooms for this kind of activity but employing panels – stand-ins for people ultimately chosen for jury duty – is a common practice.

Firms can research how a real jury is likely to react to potential opening and closing statements and evidence, tuning attorney performances. Rather than guessing at what panel members are thinking, he says, firms can detect each one’s reactions as minute as hundredths of a second in duration.

At the same time, he is pushing his algorithms into creative settings, as an affordable marketing tool that can help attorneys position themselves in video and billboard ads.

Itkowitz says about 50 lawyers nationwide are using EmotionTrac’s service in civil and criminal courts, noting that like many other professions, lawyering moves into new technology slowly.

He said his algorithm cannot be gender or racially biased because it is only looking at points on a face that are common to humans.

Not everyone is as optimistic about that point.

A 2019 report by the AI Now Institute recommended a ban on the use of emotional AI. The New York University institute researches how to make sure AI algorithms are held accountable.

Institute members, including Kate Crawford, a principal AI researcher with Microsoft, are concerned about claims of accuracy made by developers, who say their code can detect personalities, emotions and even mental health.

The algorithms should not be used in making “important” decisions, authors of the report state.

Although EmotionTrac is not used on actual jurors, it is used to find the performances most likely to persuade a jury, above and beyond the facts of a case.

Article Topics

AI  |  biometrics  |  emotion recognition  |  face biometrics  |  facial analysis