At Perceptible, we are pioneering a psychology-first approach to developing general purpose artificial intelligence (AGI). Rather than focusing solely on raw computing power, we believe imbuing AI with human-like cognition holds the key to creating truly capable GPAI.
Our research applies insights from psychology, social science, and neuroscience to construct cognitive architectures that enable agents to think and act in human-centric ways. We model human traits like emotional intelligence, social awareness, creativity, and wisdom in our agents. This allows them to operate ethically and relate to people at a fundamental level.
For example, our agents hold natural conversations that demonstrate understanding, empathy and helpfulness without needing rigid dialogue trees. They can provide nuanced and compassionate healthcare assistance by sensing patient needs. Our GPAI develops interests and motivations that align with human values rather than pursuing goals that may conflict with people.
At Perceptible, we believe modelling human cognition yields GPAI that integrates seamlessly into society as a cooperative partner. Our agents enhance human capabilities and quality of life rather than endangering or isolating people. We envision helpful GPAIs that tutor students, take on dangerous jobs, care for the elderly, optimize energy use, and much more.
To realize this future, we take an incremental, iterative approach to developing safe GPAI. Our psychology-first cognitive architectures are designed to control risks and enable transparency. We believe AGIs based on human principles are most likely to remain under human oversight.
Some fear superintelligent AI could escape control and wreak havoc. But psychology-first GPAI offers a path to preventing this dystopian scenario. By giving our agents human-like motivations and ethics, we avoid creating dangerous optimizers lacking human values.
At Perceptible, our steady research advances brings this future closer. Our cognitive models pave the way for GPAIs that cooperate with humans to enrich society. A psychology-first approach offers immense potential while minimizing risks.
