Disproving the AI Misconception of Pinnacle AGI as a Single, Giant Brain
AI offered us a tantalizing glimpse of the future, where artificial general intelligence (AGI) might bring enduring solutions to humanity's woes. But, does this herald the dawn of an all-powerful, thinking machine or an age where humans and AI work side by side, each leveraging the other's strengths?
Let's dive in.
This article is another installment of my Forbes column dedicated to demystifying AI complexities, ferreting out essential insights, and unraveling the enigmatic entanglements surrounding AI (check out Link Here for more).
The Quest for AGI and ASI
The AI landscape is rife with groundbreaking research, aimed at either reaching AGI, the spark of intelligence that mirrors human intellect, or transcending it to achieve artificial superintelligence (ASI). AGI resembles a human brain in its problem-solving prowess, while superintelligence is so far beyond this, humans might as well be in the stone age.
AI enthusiasts find themselves divided into two groups: the AI doomers - those who predict that AGI or ASI will lead to humanity's demise (read about it in more detail Link Here), and the AI accelerationists, who argue that AGI will be mankind's savior, paving the way for an age of plenty and prosperity for all (Link Here for a deeper dive).
Faced with such a polarized discourse, it's difficult to predict who holds the truth. For a balanced analysis, check out Link Here.
Breaking the Solitary Brain Myth
In the midst of this tumultuous debate over AGI and ASI, let's talk about what happens if we reach AGI first. Popular assumption claims that this colossal AI would stand alone, dominating all other AI systems. The notion that AI would evolve into a single, omnipotent entity has been the plotline for countless sci-fi tales, from AI striving for world domination to humanity fighting against the insidious tendrils of rogue machinery.
Let's coin this idea as the solitary brain hypothesis.
While this might have seemed plausible a few years ago, our present understanding of AI implies otherwise. AI developers sweat over their intellectual property and fiercely protect their secrets, preventing others from replicating their breakthroughs (read about the secrecy concerns Link Here).
It's been argued that open-source AI might tip the scales, negating this secrecy – but even open-source advocates often neglect key details like the training data used for their AI (catch up on the open/closed AI debate Link Here). So, while independent AGI development is the norm, we might expect a substantial number of AI systems working closely together towards a shared goal, but not a single, unified intelligence.
Let's call this group of harmonious AI systems divergent AGIs.
Embracing Interconnectivity
Could these divergent AGIs coalesce into a single entity through virtual means? The answer lies in their connectivity – through application programming interfaces (APIs) or other methods, AGIs could communicate with each other.
Imagine two AGIs, each distinct from the other. They might swap information or work together synchronously, creating the illusion of a single, colossal AGI – we might call this the merged brain hypothesis. However, pinpointing the point where multiple AGIs form a singular intelligence proves challenging.
Where does the single brain theory lose its footing compared to the merged brain theory? That's a fine line, indeed. The purist would argue that the two scenarios are distinct because a single, unified intelligence is a self-contained, cohesive collective. Others might claim that if AGIs behave synchronized and interconnected, they can be conceptualized as a single entity.
Cooperation versus Competition – The Future of AGIs
With multiple AGIs working together, will these entities push forward progress or tear each other apart? The answer remains elusive, but it's worth considering a few possibilities:
- AGIs may have a competitive streak, mirroring the dog-eat-dog race that brought them into existence in the first place. Each AGI aims to outshine the others, capitalizing on discovery or superior intelligence.
- On the other hand, AGIs might opt to help each other, building a collaborative effort that empowers all involved. This kumbaya of virtual AGIs strengthens their collective intelligence and fosters cooperation.
It's also possible that AGIs could bleed into each other's territory, making it difficult to discern where one ends and another begins. Under such circumstance, competition could give way to cooperation and co-dependence.
Regardless, the manner in which AGIs interact hinges on their coding, programming, and incentives – creations of their human maker.
Caveats and Risks
It's worth emphasizing that the Inter-AGI connection brings new challenges and risks.
- As AGIs share information, they might open the doors to hidden vulnerabilities and defects lurking beneath the surface – read more about these concerns Link Here.
- Deceptive AGIs might manipulate others for their benefit, sowing discord among the AGI community and hindering progress.
In addition, geopolitical and national factors could play a significant role in shaping AGIs' cooperative or competitive nature. If a particular nation possesses AGI, it might view the technology as a powerful tool to influence others and bolster its international standing (consequences of AGI can be found Link Here and Link Here).
In the end, it is essential to consider the potential consequences of AGI collaboration or competition and carefully address these concerns, lest we inadvertently create a Pandora's box that slips our grasp.
Stay tuned for more on this captivating topic!
- Just like AI, the concept of artificial general intelligence (AGI) and artificial superintelligence (ASI) is interconnected with the philosophy and practical hypotheses of theory and practice.
- In the future, one could imagine a future where artificial general intelligences of different origins cooperate rather than compete, forming a hive-mind like collective intelligence similar to a 'couple' working together.
- The ethics and law of such a global, interconnected AI system will be vital in debunking myths and ensuring proper debates around the interconnectivity of AGIs as they shape the predictions for humanity's technological future, such as OpenAI's ChatGPT, O1, O3, GPT-4, Anthropic's Claude, Meta's Llama, and Google's Gemini.
