Measures of emotion: Cognovi uses artificial intelligence to understand how we feel
We all think that we make decisions according to facts and behave logically, but do we?
Some fields of psychology and philosophy argue for a conscious, rational, objective mind. But in real life, logical fact-finding can lead to deductions and inferences that are subjectively evaluated according to a person’s belief system, the source of the information, how it’s delivered, and the anticipated effects on the person or those they love.
For the most part, we make decisions and behave depending on how we “feel” about the information we collect. We know the reasons why we shouldn’t have that drink, but it makes us happy. We know the reasons why we should go to the doctor with our symptoms, but we’re afraid of the diagnosis.
In other schools of thought, logic has nothing to do with it. Opinions, decisions and behaviors are based on complex mental and affective processes that operate out of awareness and certainly not based on logic. Whether these processes are described as intuitive or implicit, unconscious or subconscious, the experienced emotion is the primal key to understanding them.
Either way, emotions are critical. In some cases, we may have reasons behind our emotional response, but ultimately emotions rule. Other times, we have no idea why we feel or act the way we do.
If this complex emotional landscape is difficult for individuals to explore and challenging for small group interactions, it may seem impossible to define for larger populations. Advertisers and political campaigns have known for decades that listing features of a product or benefits of a candidate’s platform will only go so far.
The real task is determining not what people know about a product or candidate, but how they feel about it. In the past, surveys were often restricted to using scales that measure only narrow dimensions, typically positive, negative, or neutral sentiment. User reports and focus groups could be limited in scope and open to subjective interpretation. And many methods can be criticized for not corresponding to real-world emotional attitudes.
Enter the most unemotional quantitative tool possible: artificial intelligence. The convergence of the most significant developments in computer science has opened the door to real insights about large-scale emotional motivators and inhibitors.
Emotion AI detects human emotion by examining verbal communication as well as non-verbal cues, such as body language, tonal and facial expression. “Big data” mining analyzes large volumes of data to extract patterns of information.
Machine learning enhances computer performance by enabling programs to develop new strategies and write new algorithms. To determine the explicit or implied meaning of speech or writing, natural language processing techniques allow a semantic analysis of verbal communication and ultimately the emotional motivation behind it.
These combined tools are applied by a pioneering new company, Cognovi Labs, to extract the significant emotional attitudes and affective patterns of large populations. Founded on academic research, their innovative methods data-mine free-flowing conversations in short-form social media texts (like Twitter posts) and identify the most relevant emotional undercurrents circulating about targeted issues.
The emotions they detect are far more detailed than positive or negative sentiment, spanning the range from happiness, joy, amusement, surprise, hope, trust, and gratitude, to fear, disgust, sadness, anger, confusion, anxiety, and depression. In contrast to intuitive or subjective opinion, no matter how “expert,” they are able to document and quantify which affective components motivate behavior and which inhibit it.
Cognovi Labs analytics produce more than enhanced static descriptive reports. They document and graphically display changes in affective patterns and trends over time and across geographical areas, and classify responses by attributes such as gender and political affiliation.
For example, their Covid Panic Index was one of the first indicators of the rapid growth of fear and anxiety and its effect on economic activity, all visually displayed on maps and word clouds of emotional words and key phrases. Their recently developed Vaccine Attitudes Dashboard illustrates changes in awareness and confidence in similar presentations.
Most important for their mission, Cognovi applies their considerable behavioral psychology expertise to develop predictive and prescriptive analytics. To predict future action, Cognovi uses a proprietary appraisal theory to understand the emotional motivations of behavior operating in a population.
How well does it work? It accurately predicted Brexit, bungled by pollsters, and the 2016 US Presidential election, which surprised almost everyone.
Proactively, Cognovi emphasizes pretesting of promotional or informative messages. As one of their blogs explains, “It is critical that the emotional impact of a narrative be tested prior to a press release in order to accurately predict the public’s emotional reaction and future behavior.”
Their Emotion Trigger Marketing (ETM) platform enables this by first determining the specific phrases and words that effectively communicate a message, one that triggers the emotive drivers of the desired response. In addition, Cognovi’s analytics prescribe the best timing of a marketing event, such as when to announce a new medicine.
It doesn’t take a marketing genius to see how these tools can help advance commercial, industrial, political or financial interests.
However, the Mission of Cognovi Labs underscores that “with a powerful technology comes significant responsibility,” and that social responsibility is paramount. Their commitment to public health and safety, counterpropaganda and mental health efforts clearly demonstrate that they walk the talk.