Recently in Lusaka, the Zambia Alliance for Agroecology and Biodiversity (ZAAB) hosted a public meeting to discuss the global advance of artificial intelligence (AI) and biodigital technologies and their impact on Zambia and Zambian farmers in particular. The programme featured expert insights from Jim Thomas, AI and Market Power Fellow with the European Artificial Intelligence & Society Fund, and Esther Mwema, a digital inequalities practitioner and Atlantic Fellow for Social and Economic Equity. Attendees included farmers, academics, CSOs, faith organisations, traditional leaders, and government representatives. The aim of the discussion was to look beyond the usual sensationalist approach to the subject of AI and address its real material, social, and environmental consequences, not just in Zambia but across the continent and globally.
The first order of the day was to demystify AI. Jim Thomas kicked off his session by emphasising that what is termed “artificial intelligence” is not, in fact, an intelligent entity. Rather, it is a type of predictive computation that makes statistical guesses based on massive data inputs. Because of this, AI is far from the marvel it is often portrayed to be. For one, the AI sector is built on ‘massive looting models’, which involve the broad theft of digital data and information, frequently without permission or payment. Furthermore, it is vulnerable to mistakes (or “hallucinations”) and can replicate the biases present in its training data. “The contextual bias on the Internet gets reproduced,” explained Jim. “And that bias can be gender bias, racial bias, or cultural bias of all kinds.”
In a broader geopolitical context, Jim stressed that the AI industry’s massive demand for data, energy, water, and minerals is fuelling a “new digital scramble for Africa”. Zambia, with its young population, rich biodiversity, abundance of minerals, and water resources, is a key part of this expansion plan. In her eye-opening presentation, Esther Mwema drew clear comparisons between the routes of modern underwater internet cables, controlled by corporations such as Meta and Google, and the infrastructure of historical colonial empires to highlight how these exploitative relationships endure.
The physical footprint of the digital world is undeniable – the energy demands of the data centres that power AI are colossal! A single hyperscale data centre can consume over 5 gigawatts of power, which is more than Zambia’s entire national capacity of 3.7 gigawatts. Such demand threatens to increase global fossil fuel use and could prompt a risky expansion of nuclear power. These data centres also consume millions of gallons of water daily for cooling, raising concerns about competition for this vital, already scarce, resource. The expansion of AI threatens to reverse progress in our struggle to mitigate climate change, and it’s just the beginning.
The push for critical minerals like copper, essential for AI hardware and the green energy transition, is expected to intensify mining pressure on the continent, especially in Zambia and the Democratic Republic of Congo, where these natural resources are plentiful. The recent “discovery” of a major copper deposit in Zambia by KoBold Metals, a company backed by Bill Gates and Jeff Bezos, is cited globally as an example of how AI can be used to facilitate this extraction. But who, ultimately, benefits from these resources?
The convergence of AI with agriculture presents a novel set of challenges for Zambian farmers. Already, “digital farming” platforms are being promoted by large agribusiness corporations to farmers as an innovative way to improve their yield. However, when critically examined, these platforms are just another way for corporations to acquire even more data on Zambia’s soils and biodiversity. By engaging with these platforms, farmers risk being locked into corporate input systems that erode indigenous and agroecological knowledge by replacing it with algorithm-driven solutions. Rather than empowering farmers, these advances further infringe on the rights of farmers.
Perhaps most concerning is the emergence of “generative AI” in biology, where new viruses and proteins are being designed by AI. AI errors or hallucinations in a biological design could have dangerous consequences, raising serious biosafety concerns. Generative AI also enables a new form of biopiracy, where AI models trained on genetic data from Zambian biodiversity could be used to generate commercial products without fair benefit-sharing with the source nation.
By the end of the meeting, participants agreed that, concerningly, these technologies are advancing faster than the policies needed to govern them. There was a strong consensus that Zambia needs robust, inclusive regulations that prioritise the interests of its people. This will entail strengthening biosafety and data protection laws, asserting national sovereignty over genetic resources and data, and conducting thorough technology assessments.
This event affirmed that protecting our interests as Zambians will require a critical and informed approach. This includes prioritising bottom-up solutions like agroecology, investing in public awareness, and ensuring that affected communities have a decisive voice in shaping the technological future of the country.