AI Part 1
Artificial Intelligence is set to create a much much more dangerous world. It will almost certainly generate further increases in armed conflict and civil unrest, globally – and it will do that in multiple and sometimes unexpected ways:
 First of all, AI-driven commercial and geopolitical hunger for rare earth and other minerals is already contributing to numerous major armed conflicts. In Myanmar (formerly Burma), the military government (which came to power in an anti-democratic military coup in 2021) uses extreme violence to try to recover rare-earth-rich areas from its political adversaries (anti-junta rebels).Â
In Madagascar, security forces defending AI-relevant rare earth mineral concessions have already killed and injured at least nine land rights protesters (and have arrested more than 75 others).
 And in the Congo (the DRC) competition for control of AI-relevant non-rare-earth mineral resources (cobalt and coltan) has led to armed conflict on a vast scale – and the terrible war there is now largely funded by the world's hunger for AI-relevant minerals. So far demand for AI-relevant materials in eastern Congo has led to at least 6000 deaths, tens of thousands of injuries, well over 80,000 rapes (almost half of the victims have been children) and over 2.5 million people forced to flee their homes.
 And elsewhere in the world, the AI-driven hunger for specific minerals is likely to intensify repression in numerous countries and enable organised crime on an increasing scale.
 There is also a major risk that tensions over AI-relevant minerals (and products made with those minerals) will further exacerbate US/China relations and contribute to any potential deepening economic (or conceivably even military) future collision between those two world powers.
In the South China Sea, competition to mine the seabed in disputed maritime jurisdictions could well lead to armed conflict between China on the one hand and Vietnam and the Philippines on the other. It could also involve the USA which has recently signed rare earth agreements with several Southeast Asian countries. Likewise – in the Congo – China and the US find themselves in competition to acquire rare earth metals.
What's more, the current US government's recent (2025 and early 2026) controversial threats to annex Greenland (a self-governing autonomous part of Denmark) have been partly driven by American aspirations to control that territory's vast AI-relevant rare earth and other AI-pertinent mineral resources. In the end, the US used those threats to (de facto forcibly) extract a deal from Greenland giving America exclusive access to many of those resources.
 There is also a substantial risk that military clashes could occur between the US and other nations over America's determination to illegally mine for AI-relevant rare earth minerals on the seabed of the Pacific. France has already condemned US actions as high-seas “piracyâ€. Certainly, deep-sea mining (driven by AI demand) could lead to future military clashes between the US and China (and/or Russia). That's because rival mining initiatives could ultimately end up being awarded seabed mining concessions for identical or overlapping seabed areas by two mutually hostile bodies – the internationally-recognised UN-backed seabed-mining-authorisation agency (the International Seabed Authority [often treated antagonistically by the USA]) and by a US government domestic agency, the National Oceanic and Atmospheric Administration (which has no internationally-recognised right to award concessions in international waters)!
 But AI will directly threaten world peace in another crucial way.
That's because it will make it vastly easier and cheaper for many more countries, worldwide, to develop nuclearweapons.Â
AI is therefore likely to cause an acceleration in nuclear proliferation and resultant geopolitical instability. It will, within the next few years, enable countries to develop nuclear weaponry in literally half the time that it has taken in the past. It will also allow countries to achieve nuclear weapons status without having the hitherto crucial required very advanced scientific knowledge and resources. What's more, in cash terms, it will cut the overall cost of creating a nuclear device from drawing-board to bomb by up to 25%. Â
 What's more, AI could compromise and undermine the practice and viability of Mutual-Assured-Destruction-based nuclear deterrence – and thus remove a key factor which has helped prevent nuclear war for the past 75 years. AI weakens deterrence because it functions in literally the bat of an eyelid – and therefore compresses political/military reaction time.
 Secondly, AI will very substantially increase the potential for civil unrest – and civil conflict.
 Mass unemployment, downward wage pressure increased poverty, mass economic disenfranchisement and the concentration of economic power (in the hands of big companies, able to more effectively deploy and exploit AI) will all continue to create an even more unequal society than now.
 That will inevitably generate very serious social tensions – and governmental/elite suppression of any overt challenges to the system.
 It is therefore very likely that, as civil unrest increases, civil liberties will come increasingly under threat from government.
 What's more, AI will economically undermine quality journalism (ie accurate information available to the public) and will enable a large increase in police surveillance capacity – both of which may degrade civil liberties and democracy.
 To make matters worse, the use of AI will inevitably lead to further concentrations of power (and enhancement of top/down control) within political parties, thus further marginalising political party grassroots members and activists and driving them into broader anti-establishment movements, potentially involved in civil unrest.
 AI is also changing the very nature of conflict and warfare. The evolving technology allows military forces to kill individuals or entire groups of people (often mainly civilians) without any individual soldier being personally responsible. The development of fully autonomous weapons systems allows armies to carry out mass slaughter without any individual person making any specific targeting decisions. Known by its critics as ‘slaughter by algorithm', these AI killer systems blur accountability and will make prosecuting individuals for war crimes much more difficult. It will also make mass slaughter much easier. Semi-autonomous systems with minimal human oversight have already been deployed in the Gaza and Russia/Ukraine wars and in Libya.
Artificial intelligence will also help non-governmental would-be mass murderers – ie non-state terrorists. That's because AI will make it easier, quicker and cheaper to plan and make biological weapons of mass destruction.
 AI is likely to help terrorists to make such weapons – specifically in the planning and design phases and in the pathogen procurement and manufacturing phases.
 It's likely that it will increasingly be used to generate and operate terrorist chatbots capable of targeting and radicalising vulnerable target individuals. Indeed AI- generated deep fake images and videos are already helping terror groups with their propaganda, indoctrination and recruitment.
 And AI will also make it much easier, quicker and cheaper for terrorists to carry out cyber warfare against government infrastructure, health services, banks (and other financial institutions), transport systems, energy provision etc. Hostile governments already use cyber warfare against the UK and many other countries – but AI will increasingly widen the threat by making it much easier for relatively non-expert lone terrorists or small groups to cause very substantial economic and other (potentially lethal) damage.
 And last but not least is the potential for an AI system to ‘go rogue'. Ultimately, as AI becomes ever more sophisticated, fully autonomous AI-controlled (ie., non-human-controlled) weapons – potentially including nuclear and other weapons of mass destruction – could be used by a rogue AI system against its human owners/managers and the societies they live in.
 There is very little limitation of what will ultimately be technologically possible. The ultimate goal of AI developers is to make AI systems that replicate human thought processes (and are even capable of developing other AI systems) – so it should not come as a surprise if, at some stage in the future, an ultra-advanced AI system itself became an ambitious, armed, aggressive and lethal adversary.
Tomorrow, in the third part of this unique five-part survey of the emerging AI threat, we will look at how AI is likely to create a vast pandemic of psychological, cognitive, behavioural, physical health and educational disasters
