Neurologyca, a company that says it’s building human-centric artificial intelligence, today announced Kopernica, an AI platform designed to interpret the broad spectrum of human emotions.

Kopernica uses multimodal inputs, such as real-time audio and video, in combination with behavioral intelligence to allow AI applications to “understand” human emotions. The platform reads subtle changes in tone of voice, facial expressions, behavioral cues and mental states for stress and anxiety.

Using these markers, the company said it expects that it can allow AI models to better understand humans in different settings by adding emotional intelligence that people expect from others.

“Today’s AI systems understand what we say, but they can’t understand how we feel,” said co-founder and Chief Executive Juan Graña.

To accomplish this, Kopernica uses a unique 3D pattern recognition modeling capability that monitors more than 790 points of reference on the human body, more than seven times more than existing market solutions, the company said. For audio, the underlying model listens beyond just words and captures patterns of speech such as tone, rhythm and cadence.

Neurologyca said, unlike other platforms that rely heavily on manually labeled public datasets, it trained Kopernica on decades of scientifically grounded neuroscience research.

Beneath the signal processing component, the model uses a deep-learning framework composed of 10 processing layers, which can evaluate up to 90 classified emotions. This allows it to determine complex states such as motivation, stress, attention and cognitive load.

“With Kopernica, we’ve created the human context layer that will empower these systems to not only capture nuanced human emotions but respond with empathy, adapt their behavior, and genuinely enhance the human-machine relationship,” said Graña.

The company said Kopernica functions as an infrastructure layer, calling it an “emotional operating system” designed to work with another large language model or AI agent or application.

All of this uses local, on-device processing and anonymized insights developed to avoid storing or sharing any identifiable user data without user consent, Neurologyca said. That allowed the company to develop its infrastructure to be focused on privacy by design.

The potential for misuse of AI that could detect emotion has long been a concern for privacy and legal experts.

“Emotional AI, if not operated and supervised properly, can cause severe harm to individuals and subject companies to substantial legal risks,” warned Lena Kempe, principal attorney at LK Law Firm, writing for Business Law Today. “It collects and processes highly sensitive personal data related to an individual’s intimate emotions and has the potential to manipulate and influence consumer decision-making processes.”

Existing global regulations, such as the European Union’s AI Act, have banned AI-driven emotion detection and prediction from policing, workplaces and classrooms, with few exceptions where medical treatment and safety are concerned.

Neurologyca said if used appropriately, Kopernica could revolutionize how people interact with AI, such as OpenAI’s ChatGPT. Emotional enhancement would allow AI agents and LLMs that adapt to human emotions, that change their tone and pacing based on user inputs; wellness apps that monitor emotional and cognitive states; and media content that can detect mood and recommend better entertainment.

Using the same underlying infrastructure, Kopernica can also be used in clinical systems to flag the early indicators of cognitive strain or stroke risk.

Neurologyca is a European-based company that recently opened its first office in San Francisco. The company said there has already been strong demand from U.S.-based companies across AI leaders, wellness innovators and infrastructure partners. Kopernica is already available to select partners, with broader availability planned for the second half of 2025.

Image: SiliconANGLE/Microsoft Designer