-min%20(1).jpeg)
AI needs to understand people, not just words…
Artificial intelligence presents myriad opportunities to teach, train, support and enhance human-to-human communication. To do this effectively it needs to be socially intelligent; capable of interpreting the subtle cues that we as humans are able to pick up on subconsciously.
This is particularly important in scenarios in which AI systems are already collaborating with people – as is the case in many industries – as well as areas in which AI could be used to improve dialogue between one person and another, whether in a therapeutic, commercial or personal setting.
When AI has the capacity to navigate social and emotional communication it will be able to not only assist with functional systems, but creative ones too, such as enhancing collaboration or negotiating the dynamics of a team. It can also support people to improve their communication with others; such as guiding people through a challenging conversation.
However, the inherent complexity of social intelligence remains a hurdle for AI systems to overcome. Even in the latest GPTs need to be asked what emotions they see in an interaction. There is still a great deal of work required for AI to meet the high bar required for it to deliver as a truly socially intelligent tool.
Interhuman AI: pioneering social intelligence for AI systems
Interhuman AI is developing a social intelligence layer that integrates with any existing AI system. It is able to pick up on verbal and non-verbal cues – analyzing everything from voice patterns to body language – allowing an AI to respond to an interaction in a way that is truly human. “And we're able to do all of that in real time,” says Paula Petcu, co-founder and CEO. “Like, as we have a conversation, detecting how you're communicating.”
Their technology goes beyond regular vision models by implementing a three-layered approach including models that also interpret and reason, therefore understanding what the users’ social signals actually mean in context.
Their moat lies in these layers of interpretation and reasoning, grounded in scientific behavioral coding research led by co-founder Line Clemmensen.
The first layer, multimodal perception, simultaneously processes facial expressions, body language, and voice tonality/prosody within conversational context. The second layer, socially intelligent interpretation, translates these raw signals into meaningful social concepts using a combination of machine learning and psychological behavioal models. The final layer, social reasoning, adapts response strategies based on these interpreted signals, enabling AI- agents, avatars or any other AI system, to respond appropriately to the subtle human cues, just like a human would do in a conversation.
Interhuman is aiming to develop the frontier social intelligence AI models that capture and understand the full spectrum of human non-verbal communication.
One of the use cases and problems that the startup is addressing is the gap in workplace training. Current training is focused on dull click-through slides and videos. It’s “super boring”, says Petcu. “Everyone is just clicking next, next, on videos and images and it doesn't stick.”
By employing socially intelligent AI, this process can be transformed into an experience that is engaging and realistic – and capable of empowering people to develop new behaviours. Interhuman AI provides AI-based role plays in which different scenarios are simulated, from sales calls to challenging conversations with an employee, to leadership skills.
But the start-up is about more than solving the workplace training problem. They see their social intelligence layer being used in applications in healthcare, sales, customer service, even robotics, “wherever human communication matters”, says Petcu. “We’re not just an application on top of what already exists. We’re building foundational technology. We’re innovating.”
Experienced founders in AI research, data science and business leadership
Paula Petcu CEO/CTO: Recently listed among the Nordic tech scene’s most influential figures, Petcu is a computer science specialist with a focus on the relationship between tech, health and wellbeing. She was previously the CTO of Brain+ and the cofounder of Health Mind Tech, two companies dedicated to exploring the way technology can improve people’s lives.
Frederik Sally CCO/CMO/CPO: A 3X founder with expertise in product, marketing, design and user experience. Sally previously worked as Chief Marketing Officer for Ackerman (now part of s360) one of Europe’s leading marketing agencies, where he helped build it into a market leader.
Prof. Line Clemmensen CSO: An expert in statistical modelling and machine learning, with a particular focus on health and life sciences. Line joined the University of Copenhagen last year after a seven-year position at the Danish Technical Institute. She has secured research grants from organisations including Novo Nordisk, the Danish Ministry of Defence, Innovation Fund Denmark and the Lundbeck Foundation.
Investment Thesis:
Paula, Frederik and Line bring together the ideal blend required for an ambitious startup – they, like us, can see that too. “Coming from two worlds is an advantage, right?” says Sally. “Because you have a different view of looking at technology.”
Petcu has over a decades’ experience working in the digital health space and has a deep commitment to explorative technology. With her previous roles she has teased out many ideas with great potential to improve lives and with Interhuman AI there are the seeds to grow something that will have commercial and social impact.
Petcu and Sally connected at the Antler residency. Sally had already been scoping out Paula’s work and could see huge value in what she was doing. He is bursting with ideas of his own, but also demonstrates humility; putting his own concepts to one side and instead choosing to collaborate and learn from Petcu – as well as offer up his unique skillset to support hers. While Sally was determined to be part of a boundary breaking AI project, Petcu was seeking a partner with an instinct for business and marketing who can elevate pioneering research-based technology into a commercially viable – and appealing – product.
“Most AI companies are fighting for attention in an overcrowded market," says Sally. "And I think we'll see a lot of them fail this year. We're taking a different path with Interhuman - focusing on a core problem that LLMs don't solve. We're building a foundational layer that significantly enhances LLM capabilities and gives developers real control over these systems.”
That they can do this with the support of Line’s leading expertise in machine learning is a further boost to the Interhuman AI operation. As they know, the AI space is moving fast – while they are ahead of the curve now, the competition is fierce. By investing time and resources into research they position themselves well to remain at the cutting edge of socially intelligent AI.
For all press enquiries: press@antler.co
