Posted in:

Peter Voss and His Vision for Artificial General Intelligence (AGI)

Peter Voss, a cognitive scientist, engineer, and entrepreneur, was born in Germany in 1954. He has made substantial contributions to the field of Artificial General Intelligence (AGI). Throughout his career, Voss has established numerous companies and remained a trailblazing figure in AGI, emphasizing a distinctive approach that he believes is essential for developing genuine human-like intelligence in machines.

In 2002, Peter Voss, Ben Goertzel, and Shane Legg introduced the term “Artificial General Intelligence” to denote AI systems that are capable of thinking, learning, and reasoning in a manner similar to that of humans but with enhanced capabilities. Voss clearly differentiates “Real AGI” from the current AI hysteria that frequently envelops large language models (LLMs). He defines Real AGI as AI that is capable of learning in real-time, autonomously completing any cognitive task that a clever human can perform, and achieving this with minimal training data, time, and resources.

Voss contends that Cognitive AI is the more viable approach to Real AGI than Generative or any other Statistical AI. He critiques the inherent flaws of current LLMs, which include their error-proneness, inability to update their models in real-time, and absence of metacognition. These systems are economically unsustainable in the long term due to their substantial data, CPU, and power requirements, which necessitate human intervention to validate and adjust their outputs.

On the other hand, cognitive AI is founded on a profound comprehension of human intelligence and is designed to replicate its adaptive, conceptual, and interactive high-level capabilities. DARPA has classified this approach as the “Third Wave of AI,” a stark contrast to the previous waves, which concentrated on symbolic representations (GOFAI) and large data-driven deep learning.

Cognitive AI endeavors to replicate the distinctive characteristics of human intelligence, including the capacity to generate abstract concepts, learn incrementally in real time, and solve autonomous problems. Voss underscores the necessity of incorporating insights from a variety of cognitive science subfields, such as philosophy, cognitive psychology, and computer science, into Cognitive AI.

Key Characteristics of Human Intelligence is defined by its capacity to comprehend the world, accomplish a diverse array of innovative objectives, and independently integrate and modify knowledge and skills through continuous incremental learning. This intelligence operates in real-time, with restricted knowledge, time, and resources. Humans are capable of recursively constructing abstract conceptions, which is crucial for the development of self-awareness and language skills.

In order for AI to emulate human cognitive capabilities, it must be capable of incrementally updating its knowledge, inferring new implications from supplementary information, and operating autonomously with metacognitive awareness. The substantial distinctions between Cognitive AI and contemporary Statistical AI models are underscored by these requirements.

Typically, cognitive AI is implemented through a “Cognitive Architecture” or “artificial brain,” which can be implemented in robots or as standalone computer programs. In order to facilitate this architecture,, the company founded by Voss, has implemented cognitive algorithms and developed a high-performance graph database. This method, which is elaborated upon in their whitepaper, exhibits substantial enhancements in performance in comparison to current systems.

In contrast to Generative AI, which has dominated recent advancements and funding, Cognitive AI has received relatively little investment despite the promise it holds, according to Voss. He predicts a change as more stakeholders acknowledge the potential benefits of AGI and the constraints of current AI models.

Peter Voss’s vision for AGI is founded on a profound comprehension of human cognition and a dedication to the creation of AI that can think and learn in a manner that is genuinely akin to that of humans. His emphasis on Cognitive AI as the pathway to Real AGI provides a compelling alternative to the current trends in AI development. Voss’s groundbreaking research continues to motivate and influence the pursuit of genuine artificial general intelligence as the field develops.