By Ashoke Agarrwal
The AI mavens have begun to look beyond the wonders of Large Language Models and Generative AI towards the goal of AGI (Artificial General Intelligence).
The question is, what is AGI?
Will it be an AI that is super intelligent and capable of everything, in other words, superhuman? Will it be a single entity or a group of competing for-profit and open-source entities? Wouldn’t logic dictate that superhuman entities would, in the end, coalesce into one single omniscient and omnipotent entity – that the Singularity is not just the development of AGI but this integration that would threaten the much-dreaded takeover by AI of human civilisation?
Alternatively, this technology could take another evolutionary path. The first step would be to alter the metaphysics of the technology, starting with the nomenclature. The term Artificial Intelligence highlights the non-human nature, and thus, from the outset, puts it in a competitive stance vis-Ã -vis humans and humanity, conflicting with the fact that this is yet another human-developed technology that will, like all other technologies, will be helpful, by and large, while being misused by some and deleterious in some contexts.
And so it has been proven to date. DeepMind’s AlphaFold will drive the development of revolutionary healthcare products, and Generative and Agentic AI will enhance white-collar productivity. Meanwhile, some will use AlphaFold to produce bio-weapons, while others will utilise Generative AI to create deepfakes that sow misinformation and distrust.
Wouldn’t it be better to name this technology Augmenting Intelligence (AI) instead of Artificial Intelligence? AuI will be technology that is joined at the hip with humans and their needs, wants, and aspirations.
Instead of the quest for AGI, the quest then becomes for the ultimate enabler of humans—the Augmenting Super Intelligence (AuSI). Creativity is at the pinnacle of intelligence, and it is the result of the marriage of high competence and the individual’s unique perspective on the world, in other words, their personality.
In the age of AI, AI will be the companion of each individual, having grown with them since childhood, with permission-driven access to the individual’s life activities – their communications, interactions with the world, and the use of special tracking devices to monitor their physical, mental, and emotional states. AuSI will enable individuals to accomplish every task, marrying its insight into the conscious and subconscious states of the individual with a vast knowledge bank and the speed of retrieval and synthesis of a digital, connected entity. As the individual grows into a profession or a walk of life, their AuSI will become their enabler of a high level of creativity. The AuSI communicate through screens, augmented and virtual reality. As technology develops, an individual can choose to integrate their physical self with technology, becoming a cyborg. Each individual’s AuSI will be under their control in terms of access and ownership. An individual’s AuSI will also promise life after death in the sense that the AuSI can become a digital entity that interacts and works with the world. Other entities, such as organisations, business entities, and governments, will have their own AuSI avatars.
Humanity’s evolutionary path has many forks. One fork is where we go extinct, like the dodo felled by hubris, by either nuclear extinction or ecocide or both. The other fork is that we go down the road of creating an omniscient, omnipotent AI that extinguishes the human race or degrades it into a servile species. The best case is a fork where we course-correct and go down the AuSI road, flourishing further as a species, graduating, as Yuval Noah Harari put it, from Homo Sapiens to Homo Deus.
PS: I have been grappling with the idea of AuSI for years now. Back at the start of 2022, much before ChatGPT informally dawned the age of AI, I wrote about it as Concierge Intelligence (CI).