As the relentless growth of artificial intelligence accelerates across the global economy, the technology is already posing profound and immediate challenges to modern society. Corporations across various sectors have steadily begun laying off human workers whose routine and complex jobs AI models can now perform with speed and efficiency. Simultaneously, the massive physical expansion of energy-intensive data centers has triggered a sharp spike in global energy consumption, significantly worsening the carbon footprint of the technology sector at a time when climate targets are already under severe strain.
Beyond economic and environmental pressures, malicious actors are increasingly weaponizing autonomous AI agents to automate the mass production and dissemination of disinformation, muddying public discourse and undermining democratic integrity. Advanced algorithms are frequently demonstrating systemic bias, raising urgent ethical concerns about flawed applications and institutional discrimination by human-related technologies that rely heavily on these opaque systems. Furthermore, AI is now actively being deployed to identify targets in active warfare, transforming the mechanics of combat and leading to a greater, deeply concerning loss of human life on the battlefield.
These are complex, multifaceted problems that demand serious, deliberate regulatory attention from governments and international bodies. Yet, paradoxically, today’s dominant AI policy conversations are rarely grounded in these immediate, tangible challenges. Instead, the global policy discourse has been comprehensively swept up by apocalyptic prophecies and speculative anxieties about the distant future.
We see this pervasive doomsday narrative actively promoted by influential figures within the AI industry itself, and echoed uncritically at the highest levels of political and institutional decision-making. From high-profile lawmakers such as United States Senator Bernie Sanders championing strict bans on theoretical superintelligence, to cautionary messages delivered by the Vatican, to former U.S. Secretary of State Hillary Clinton publicly labeling open-source AI as a genuine and present security threat, the focus has shifted dramatically away from the material present.
It is entirely reasonable for global leaders to take warnings originating from the tech industry seriously. However, this powerful chorus of influential voices sounding the alarm about nebulous "existential risks" has inadvertently pulled public and regulatory focus away from the multiple ongoing crises that AI has already caused, which are actively affecting labor markets, local environments, and global communication channels.
Worse still, this manufactured public panic is increasingly being weaponized by powerful actors to enforce a modern regime of digital colonization—defined primarily by the alarming concentration of economic and technological power in the hands of just a few dominant AI corporations based in a handful of wealthy nations.
The Weaponization of Safety and the Attack on Open-Source AI
The prevailing doomsday narrative provides the absolute, perfect justification for regulatory frameworks that disproportionately target open-source AI models. Open-source models are typically free or significantly cheaper to use, and crucially, they allow everyday users and smaller organizations to modify them to fit their specific local needs and cultural contexts. The political logic presented to lawmakers by tech lobbyists is simple, seductive, and ultimately misleading: if advanced artificial intelligence is an inherent potential weapon of mass destruction, it follows that it can only be safely developed, controlled, and deployed by a trusted few corporate entities.
Under the benevolent banner of "safety," major technology monopolies are actively attempting to write the regulatory rules of the road to protect their market dominance by legally shutting down or heavily restricting competition from smaller, more affordable open-source AI models. This regulatory capture directly mirrors historical colonial tactics, where dominant empires systematically prevented local industries from developing in order to force total, long-term reliance on the empire’s expensive manufactured exports.
Currently, the entire global community is heavily dependent on frontier AI systems developed by a handful of tech giants. This vulnerability is felt most acutely across the Global South, which largely lacks the astronomical financial resources required to create its own frontier AI models completely from scratch. Developing such systems now costs upwards of one hundred million dollars in raw compute power alone, placing them entirely out of reach for developing nations and local entrepreneurs.
The fundamental problem with relying exclusively on these dominant, imported AI models is not merely that they are prohibitively costly to license and run. Crucially, they routinely fail when applied to local, real-world contexts that differ vastly from the Silicon Valley environments where they were conceived.
A striking example of this disconnect can be seen in Kenya’s recent rollout of the AI-driven Social Health Authority (SHA) system. Because the imported, off-the-shelf model was built on Western assumptions that failed to accurately capture the economic reality of Kenya’s vast and complex informal economy, the system failed spectacularly in practice. It systematically overestimated citizens’ actual incomes and consequently overcharged the country’s poorest citizens for essential healthcare premiums, directly harming the very vulnerable populations the reform was intended to help.
Preserving Local Sovereignty Through Open-Source Innovation
To be genuinely effective and equitable, frontier AI systems must be carefully adapted to local data, languages, and cultural frameworks. Because building foundational models from scratch remains financially impossible for the vast majority of nations, open-source AI serves as the only viable bridge for local innovators, researchers, and public institutions. More importantly, preserving the open-source ecosystem allows the Global South to step off the relentless Western treadmill of human-mimicking chatbots and consumer-focused gimmicks.
Developing nations do not necessarily need conversational AI systems designed for Western entertainment or office productivity. Instead, they desperately need specialized artificial intelligence that can help tackle pressing domestic development challenges in critical areas such as energy access, public healthcare, modern agriculture, local manufacturing, and climate mitigation. If we crush open-source AI out of a disproportionate, misplaced fear of runaway AI agents, developing nations will lose the primary mechanism they possess to build their own sovereign technological ecosystems for sustainable development.
Furthermore, these emerging AI monopolies effectively lock everyday users into a relentless cycle of data extraction and severe cognitive dependency. We are already witnessing the early, troubling stages of cognitive degradation as users increasingly outsource their critical thinking, creative agency, and independent problem-solving skills to proprietary frontier AI systems.
This technological dependency creates a self-reinforcing feedback loop: our growing reliance on automated systems fuels the continuous collection of more personal and institutional data, which in turn trains even more capable models that further strip away human agency and centralize absolute control in corporate boardrooms.
While the Global South is uniquely and severely vulnerable to these dynamics due to longstanding historical inequities, this dependency ultimately traps anyone who is not sitting at the technological frontier. Even wealthy nations located in the Global North face the exact same emerging threat of vendor lock-in, proprietary control, and widespread data extraction.
Charting a Balanced Path Forward
Faced with these sweeping structural challenges, what is the appropriate solution? Demands for sweeping AI bans are certainly not the answer. Completely abandoning artificial intelligence would mean walking away from a technology that is uniquely positioned to accelerate the discovery of life-saving drugs and innovative solutions to the global energy and climate crises. There is already tangible evidence of this immense potential demonstrated by scientific tools such as AlphaFold and MOFSeq, which are transforming biochemical research.
Similarly, banning or heavily restricting open-source AI models is a deeply flawed choice. True safety does not require sacrificing equitable access to technology. We can—and most certainly must—fund robust safety research and implement rigorous testing protocols to guard against near-term security breaches and malicious use. However, we cannot allow the paralyzing fear of hypothetical, distant futures to justify the deliberate, present-day consolidation of corporate power.
Safeguarding the future of artificial intelligence means ensuring that its ongoing development is decentralized, transparent, and globally distributed rather than concentrated in a few private monopolies. It means defending open-source innovation as a vital public good and a fundamental weapon against modern digital colonisation.
It means proactively addressing the human cognitive toll of automation, ensuring that artificial intelligence serves to augment and elevate human intelligence rather than erode it. Finally, it means anchoring absolute legal and moral accountability firmly in human hands, instead of anthropomorphizing software agents that possess no humanity or conscience of their own.
The views expressed in this article are the author’s own and do not necessarily reflect the editorial stance of the publisher.