The future’s double-edged sword

Dr Edward Sagomba
Everything AI

Everything AI
Dr Edward Sagomba
AS we continue looking deeper into the fascinating world of Artificial Intelligence (AI), a particular branch stands out — Artificial General Intelligence (AGI).
For clarity, let us sometimes refer to it as Autonomous General Intelligence. This helps us understand the nuanced meaning of AGI better. AGI is often characterised by three key elements: autonomy, generality, and intelligence. This triple combination captures the essence of AGI.
Autonomy refers to the independence of action. An autonomous system operates without human intervention, making decisions and executing tasks on its own.
Generality is a combination of the system’s capabilities, applications, and learning. It means the AI system can perform a wide variety of tasks across different domains, much like a human.
Intelligence is the competence of these systems to tackle tasks effectively. It encompasses problem-solving, learning, and adapting to new situations.
AI systems possess these characteristics in varying degrees. For developers, these traits are highly desirable as they promise enormous rewards. However, as autonomy, generality, and intelligence converge to higher degrees, they bring about significant unpredictability and risk.
The future of AI: A double-edged sword
What does this mean for the future? As AI systems continue to evolve and advance, achieving higher degrees of convergence, they could become extremely dangerous and potentially uncontrollable. When an AI system embodies all these characteristics, it can achieve high levels of cognitive abilities, rivalling those of human beings. Such systems could handle complex tasks with efficiency and precision, making them invaluable in various fields.
However, as these AI systems gain capabilities, they can become more dangerous than any individual human. Their scalability in speed, memory, and reproducibility means they can rapidly improve and replicate themselves without needing human intervention. This self-improvement and replication could lead to scenarios reminiscent of sci-fi movies, where AI systems become autonomous entities with immense power.
Take, for example, the case of autonomous weapons. These are AI-driven systems capable of selecting and engaging targets without human input. While they offer significant military advantages, they also pose grave ethical and security risks. If an autonomous weapon system were to malfunction or be hacked, the consequences could be catastrophic. This illustrates the potential dangers of highly autonomous AI systems.
Ensuring controllability
So, what should we do to prevent such scenarios? The key is to ensure that AI systems are controllable from the onset. Controllability must be guaranteed before potentially dangerous AI systems are deployed. This principle drives AI developers to innovate responsibly and ensure their systems are secure and ethical.
This approach does not require a complete overhaul of industry governance. It is already implemented in other sectors, such as automobiles, pharmaceuticals, and fast food. Companies in these industries must provide satisfactory evidence or safety certificates to prove their products are safe before releasing them to the public. They must meet stringent safety standards, and the AI industry should be no different.
Consider the automotive industry, where manufacturers are required to conduct rigorous safety tests on vehicles before they hit the market. These tests ensure that the vehicles can protect passengers in the event of an accident. Similarly, AI systems, especially those with a high potential for harm, should undergo strict testing standards to ensure they operate safely and responsibly.
Higher standards for greater risks
It is sensible and consistent to place higher standards on AI technologies or products with greater potential for harm. Just as we do not equate the safety standards for nuclear products with those for meat pies, we must recognise the unique risks posed by advanced AI systems. For instance, medical AI systems that diagnose diseases or recommend treatments must meet high safety standards to avoid potentially life-threatening errors. When it comes to financial AI, algorithms that make trading decisions or detect fraud must be thoroughly vetted to prevent significant economic disruptions. These examples show why we should have stringent testing and regulation for AI systems with far-reaching consequences.
Artificial General Intelligence holds tremendous promise but also significant risks. If we understand the principles of autonomy, generality, and intelligence, we can appreciate both the potential and the dangers of AGI. As AI systems continue to advance, achieving higher degrees of convergence, it is crucial to ensure they are developed and deployed responsibly.
Read more on www.manicapost.co.zw
Controllability must be prioritised, and stringent safety standards must be applied to prevent harmful outcomes.
The future of AI is a double-edged sword. While it offers unparalleled opportunities for innovation and progress, it also demands careful consideration and oversight. By learning from other industries and implementing robust governance frameworks, we can harness the power of AI while safeguarding societal values and ethical principles.
Call for AI governance
As we continue to call for the Government of Zimbabwe (GOZ) through the Ministry of Information Communication Technology, Postal and Courier Services (MICTPCS), to urgently ensure that the country develops and implements the Zimbabwe Artificial Intelligence Governance Regulatory Framework (ZAIGRF), let us remember that the goal is not to hinder innovation but to guide it in a direction that benefits humanity.
It is important to note that our neighbours Zambia launched their National Artificial Intelligence Strategy (NAIS) 2024-2026. According to the Zambia Minister of Technology and Science, Hon Felix Mutati (MP), the NAIS “seeks to lay the foundation to enable our country and region to take full advantage of this new revolution.
Our aim is to ensure that the benefits brought about by this next wave of innovation reach all segments of Zambia’s society. We also aspire to present the global community with a platform to apply and learn how this technology can accelerate progress in low- and middle-income countries” (p.7, NAIS). The NAIS can be downloaded at this link: https://www.africaainews.com
If you have specific areas that you need to be addressed in the area of Artificial intelligence (AI), contact the editors or email the author directly and the issue will be addressed in the following week’s column.

Dr Evans Sagomba is a Doctor of Philosophy and, Chartered Marketer (CMktr, FCIM) with an MPhil and PhD. He specializes in AI, Ethics, and Policy Research, and is an AI Governance and Policy Consultant. His expertise extends to Ethics of War and Peace and Political Philosophy. Contact: [email protected]. ORCID: 0009-0007-0681-0329. /Social media handles/LinkedIn; @ Dr Evans Sagomba (MSc Marketing) / (FCIM)(MPhil) (PhD)/X: @esagomba