| Director of AI at arm | Award-winning AI Researcher | Best Selling Author | Strategy and Tactics | Fellow at the AI Fund | Advisor to many | Inspiring the world about AI | Contact me! |
132,035 followers
Artificial General Intelligence (AGI) remains a nebulous and contentious concept within the AI community. Despite its growing prominence as a goal among major AI companies and its portrayal in popular media, we lack a clear, universally accepted definition. Ask five different people, and you’ll get ten different answers. This lack of clarity poses a significant challenge to our progress towards achieving AGI. The term ‘AGI’ is often perceived as a goal without a solid scientific foundation, a crucial aspect we must address. The term gained traction around 20 years ago but has, in my opinion, become more of a buzzword than a well-defined objective. At least in the fears of the many, I observe a conflation between AGI and Artificial *Sentience*. However, the current trajectory of AI research, which focuses on transitioning from narrow AI to broader AI, barely scratches the surface of what would be required to create artificial sentience. Enhancing intelligence alone does not equate to creating an entity capable of sentient experiences, and true intelligence isn’t exactly linear. You don’t plot a line of intelligence; and then — beyond a certain point, call something alive or sentient. Ongoing research into growing the intelligence of a system towards superhuman capabilities isn’t a route to a judgemental, all-encompassing AGI that the science fiction movies of the world tell us will kill us. Indeed, we have already seen computers be super-human in many scenarios, and we’re still around! :) True AGI would need to encompass advanced cognitive abilities and a deep understanding of and interaction with the world, akin to human consciousness. Moreover, the quest for AGI involves far more than developing smarter algorithms; it requires a profound leap in our understanding of consciousness, emotions, and self-awareness. Current advancements in AI, such as large language models, represent incremental progress rather than revolutionary steps toward artificial sentience. These systems, while impressive, operate within the confines of their programming and lack true understanding or subjective experiences. They simulate that through language, and we infer they are ‘thinking’ or ‘reasoning’, but it’s artificial. In conclusion, the realization of AGI is a distant dream, primarily due to our struggle to define it precisely and the vast disparity between current AI capabilities and the intricate nature of human-like intelligence and sentience. As the New Scientist article rightly points out, AGI remains more of a speculative concept than an imminent reality, and our journey toward it is just beginning. And perhaps this is why companies are disbanding any safety or research around the potential impact of AGI -- because, for now, at least, perhaps it's a fallacy to do so. #ArtificialIntelligence #AGI #Futurism