Superintelligence: Bring Me the Tea, David

Quite recently, an AI researcher working with Anthropic, predicted that “there is a greater than 10% chance artificial intelligence could cause human extinction within the next decade” (The Guardian, September 2026). This speculation triggered a warning across industry and governments highlighting AI’s shortcomings and potential misuse. Nevertheless, in other news, the malicious use of AI and nuclear conflict posed the biggest threat to human extinction by the end of this century (Sky News. October 2026). The loss of life because of a nuclear war is inevitable, but this raises the question “Can superintelligence agents eradicate the human populace?” In fact, both scenarios brilliantly demonstrate ‘human stupidity’ noting how far we’ve come with technology, and yet, how far we have taken several steps back in our judgement.

A Nervous Investor Market

AI is often misunderstood by marketeers, academia, and consumes. Likewise, science fiction writers and filmmakers typically depict a rogue machine bringing humanity to its knees. Let’s not forget, that AI isn’t a technology rather it’s a concept a “What if…” if you prefer. For now, AI tools are simply clever software and smart technology or ‘assistive technology’ where such tools provide a solution for the mundane. “Technological unemployment” is a term often used to describe the inevitable job loses with the introduction of technological changes where AI tools replace human labour.

If a machine begins to think, it is no longer a machine.

 We regularly see or hear various interpretations of what AI is and what it might achieve. AI is undergoing a never-ending hype-cycle, while the Nasdaq, during the midterm US election suffered. a fall of trillions causing a nervous investor market (Reuters, November 2026) which potentially leads toward an ‘AI winter’ (BBC News, January 2020)

Artificial Intelligence

More so, the following descriptions have been compiled from opinions, industry and academia, along with a soupçon of my own thoughts. AI tools mimic human-like cognition in a machine using software, as proposed by. Dr John McCarthy et al. in his 1955 research project (A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence). McCarthy is recognised as the true ‘father’ of artificial intelligence, while earlier projects such as the ‘Turing Test’ (Computing Machinery and Intelligence, Alan Turing, 1950) has been categorised retrospectively as AI.

 We have an uncensored belief in something we have not entirely understood.

 The “Oppenheimer moment” is a term used in artificial intelligence, which compares the rapid creation and associated fears of today’s AI systems to Dr Julius Robert Oppenheimer’s development of the atomic bomb during the ‘Manhattan Project.’

Artificial Intelligence is an umbrella term—it does not understand or think and is void of consciousness. Typically, at this base state, AI performs functions that may need human intervention and surpasses rudimentary tasks, such as voice assistance, spam emails, chatbots and other primitive functions. After all, machines can execute functions much quicker than a human, plus the AI tool doesn’t need a cigarette and coffee break.

• Generative AI

  A subset of AI, where some AI tools formulate its instructions based upon good quality data collated from its environment, for example. Machine/Deep Learning (ML/DL) and Neural Networks (NN)s, along with Large Language Models (LLMs) techniques are used to aid AI tools to foretell patterns which, in turn, generate new content. The same tools may require a ‘prompt’ presented by the user, which is parsed in an attempt for the tool to ‘understand’ what you have requested. The tool attempts to mimics human creation. If the prompt is well-balanced, accurate and hardy; AI tools such as ChatGPT, MidJourney, and Claude, for example, can create text, images, audio or code. These tools are extensively used and continues to mature.

• Agentic AI

  A subset of AL, Agentic AI systems are widely used and can accomplish tasks with no or minimal human supervision. The Agentic AI ecosystem comprises agents using ML models to replicate human decision-making to complete tasks in real-time. Likewise, in a multiagent system each agent is assigned a specific problem to solve. Such agents are managed and coordinated with an orchestration model. In short, AI agents and their integration, manages the deployment, implementation, integration and maintenance of the agents.

• AGI

  Currently, Artificial General Intelligence (AGI) remains theoretical, although research continues to help better understand its potential. A subset of AI, it’s envisioned that AGI will achieve a level of human cognition that may succeed most humans. Moreover, it will also be able to learn, reason, and complete most intellectual tasks without specialised training and human intervention. This wholly differs from superintelligence because I would argue that AGI is a stepping stone toward superintelligence.

Superintelligence

The tern ‘superintelligence,’ for me, alludes to a digital being (physical humanoid) or system that may surpass the intellectual acuity of humans in all domains. It’s a theoretical future and does not exist today. Nonetheless, with the accelerated development of AI systems, I don’t see superintelligence sitting beneath the umbrella of AI; rather, superintelligence is a standalone classification with its own subsets. A superintelligent agent doesn’t persist as artificial intelligent since the agent is an ‘artificial lifeform with intelligence.’ This is humanity’s ‘Playing God’ moment with the creation of new life. As for the timing of these agents appearing in human society, I’m inclined to suggest that this is at least a few generations away.

 When does an algorithm become consciousness and mind?

 A superintelligent agent may subsist without human form where its existential presence survives within a closed environment, such as a physical machine, but has access to the internet. Naturally an agent may also exist with human form that resides in an open environment where the agent can interact with humans face-to-face. More crucially, human nuances such as facial features and dexterity should appear in human-like form. This would provide confidence with agents performing roles in health care, care givers, and other roles where humans need the reassurance that the agent is much more than a robot.

Superintelligence is an existential agent that possesses the intellectual capacity to think, debate and make its own decisions; to reason, rationalise and question its own existence.

• Policing and Self-governance

  It’s astonishing to read that the ‘big tech’ giants will police and self-govern their own AI development where the definition of ‘intelligence’ is changeable from one company to another. The fear of such a state is the loss of control between companies. A global understanding must be agreed such that companies may share their progress, development and testing milestones. Despite growing concerns regarding superintelligence, and its future, there will inevitably be someone developing such agents with malevolent intent. As such, an agreed independent agency with no affiliation must be able to internally inspect code, testing results, its abilities and future Roadmap.

• The Loss of Control

  As creators, we must ask ourselves, “How much intellect and volition should we bestow our superintelligent agent?” and “Should we share his Divine Fire?” In Greek mythology the ‘Divine Fire’ symbolises life, the creation of civilisation, intelligence and technology. It also provided a boundary between the gods and mortals.

  The Greek god Zeus kept the fire hidden from the mortals leaving them vulnerable. It wasn’t until the Titan Prometheus, stole the fire and gifted humanity enabling them to nourish, make tools, stay warm and create civilisation. The fear of gifting superintelligent agents’ autotomy is loss of control, yet we do have the ability to switch off and on features within our agent.

Until Next Time...

Our desire to achieve superintelligence has been fictionalised in many sci-fi movies and inspires us to attain such beings in human society in the belief their creation will help humans in many capacities. In the Alien franchise, for example, we have seen so-called ‘synthetics’ that positively or negatively aided the crew. Synthetics, for want a better term, could become our benchmark for their appearance in society serving several roles and functions.

In the opening scene of the film, “Alien: Covenant,” 2017, both Peter Weyland (Guy Pearce) and David (Michael Fassbender) are engaged in a heated debate. David asks Weyland who created him where Weyland rejects any random chance. David flaunts his superior intellect and questions his own immortality and discusses Weyland’s inferior biological limitations and his ‘memento mori’ (the inevitability of death). To cease David’s continued questioning, Wayland reasserts his authority as a ‘father’ and harshly commands David to bring him tea.

David as portrayed in the film is my ideal superintelligence agent not including the narcissistic and malicious agenda. His primary role is to navigate the ship and to ensure safe keeping of the ship and its crew.  As I mentioned earlier the agents working in human society is many generations away. So raising the question will superintelligent agents eradicate humanity is subject to the creators and their agenda, as there will be some companies that are just untrustworthy.   

So, this is where an "AI philosopher," Dr G signs off.