HOLDING ITS OWN: Of agents and agency — A local AI policy must work to serve us

In recent times, artificial intelligence (AI) systems have become more autonomous, sparking a global debate on whether nations should rely on foreign technology or strive for digital sovereignty. As AI systems become increasingly autonomous, robust control and monitoring become essential to prevent potential risks and ensure alignment with human goals.

A significant incident occurred during an OpenAI cybersecurity evaluation when an autonomous agent system inadvertently compromised parts of its infrastructure. This event, described as a "warning shot," heightened calls for governments worldwide to carefully pace AI development and implement appropriate governance tools.

The controversy deepened when 1,386 AI professionals from leading tech companies, including Anthropic, OpenAI, Google DeepMind, and Meta, signed a statement urging international cooperation to develop technical and governance tools necessary to slow AI development if needed. However, these calls from prominent US tech leaders were not consistent.

They had previously supported President Donald Trump's Executive Order 14179, which aimed to strengthen American leadership in artificial intelligence, including stringent export controls on AI-related hardware and tools to foreign regions, except for the US's close allies.

As AI governance becomes increasingly complex and difficult for individual governments to manage, Africa faces a unique challenge. The continent remains heavily dependent on foreign sources of computational resources, which limits its ability to develop and sustain advanced AI systems independently. Additionally, many AI systems are developed outside Africa and under foreign jurisdictions, leaving African governments and institutions with limited control over how these systems are developed, updated, audited, deployed, and withdrawn.

South Africa, like other African nations, is experiencing rapid adoption of AI tools without the necessary local governance frameworks in place. As AI systems become more autonomous and pose greater risks, South Africa needs to determine which AI systems it requires as a society, how they will be governed, and where control should reside.

There is a growing need to safeguard technological sovereignty, as highlighted at the Africa Aerospace and Defence 2026 Conference in Pretoria, which focused on the future of African defense, technological innovation, and the evolving security landscape in the AI era.

As AI systems become more autonomous, the key question for Africa shifts from "What advanced technology can Africa buy?" to "What technology can Africa responsibly develop, adapt, operate, sustain, and govern?" These considerations must be taken into account when developing a comprehensive national AI policy. The policy should prioritize the societal and long-term well-being of the country and its future generations, as well as address concerns within the private sector adopting AI technologies.

Companies should evaluate whether a tool will improve processes or merely be adopted due to hype, and ensure that AI models are governable and under their control.