South Africa’s Iris classroom robot has become a striking symbol of the country’s ambitions in artificial intelligence, education and locally driven innovation. Yet behind the excitement lies a more complicated story involving funding, manufacturing, classroom readiness and questions about scale. Launched in KwaZulu-Natal in August 2025, Iris was presented as a pioneering educational technology capable of supporting learners across the country’s 11 official languages. SABC News report on Iris
Iris was developed by Bonisiwe “Thando” Gumede, a former educator who moved into technology after recognising the potential of software to reduce some of the administrative pressures confronting teachers. The robot combines a mobile humanoid platform, interactive display and voice capabilities with educational software. It is designed to answer questions, explain lessons, assist with assessments and provide repetitive instructional support, rather than remove teachers from classrooms altogether.
The technology’s most distinctive selling point is its linguistic reach. Iris has been developed to operate across South Africa’s 11 official languages while being aligned with the national curriculum from Grade R through matric. That ambition matters in a country where language can strongly influence educational access. A digital assistant that can interact naturally with pupils in their preferred language could make technology more inclusive, especially in schools with limited specialist educational resources.
However, the distance between a promising demonstration and nationwide deployment is substantial. Recent reporting indicates that only five KwaZulu-Natal schools currently have Iris systems operating in their classrooms. Those installations were provided by Gumede’s company, BSG and Technologies Services, rather than through a large government procurement programme. The gap between the original nationwide ambition and the present footprint highlights how difficult it can be to move an innovative education project from publicity into sustained public-sector implementation. TechCentral investigation into Iris
Cost is another major obstacle. A complete Iris installation is reported to cost approximately R650,000 per school, incorporating the robot alongside computers, display equipment, connectivity and cybersecurity infrastructure. Multiplied across South Africa’s public-school network, the financial requirement becomes enormous. TechCentral calculates that full deployment could reach approximately R14.5-billion. Such figures inevitably raise questions about whether expensive physical robotics represents the best use of limited education technology budgets.
Government support has also proved more complicated than the launch headlines suggested. The Department of Science, Technology and Innovation has publicly backed South African innovation and STEM development and participated in events associated with Iris. However, the department has confirmed that it has not provided funding for the project. This distinction between institutional endorsement and financial commitment is crucial. Education is largely administered through provincial structures, while national innovation agencies have different mandates and funding responsibilities. Department-backed Iris launch coverage by SABC News
There is also an international dimension to Iris that deserves attention. Gumede developed the educational concept in South Africa and subsequently worked with Indian robotics company MakerLabs. Components are being prepared in India and assembled in Durban, while the South African team continues working toward greater domestic production capability. The project therefore illustrates both the possibilities and limitations of international technology partnerships: local intellectual and educational expertise can combine with overseas robotics capabilities while domestic manufacturing capacity develops progressively.
The question of whether a physical robot yields superior educational outcomes compared to software alone remains unresolved. Research into educational robots suggests that they can be useful when they complement teachers, particularly by providing interaction, repetition, and engagement. But researchers have also warned that the novelty of a robot can diminish over time. This makes long-term classroom evidence more important than impressive demonstrations. Iris will ultimately need to prove that its educational benefits justify its hardware, maintenance, and connectivity costs.
The project also raises a broader question about sustainability. A classroom robot is not simply a piece of equipment purchased once and forgotten. Schools require software updates, technical support, replacement components, connectivity and teacher training. International experiences with educational robots have demonstrated that commercial or technological changes can leave schools with expensive equipment that becomes difficult to maintain. Iris therefore needs a durable support ecosystem if its ambitions extend beyond a handful of demonstration classrooms.
Perhaps the most revealing part of the Iris story is that the robot is competing with cheaper forms of artificial intelligence. Other South African education initiatives are using mobile and messaging platforms to deliver AI-supported tutoring without requiring schools to purchase specialised machines. That does not make Iris irrelevant; its physical presence could encourage interaction and STEM awareness in ways a smartphone application cannot. But the central test is practical: can Iris deliver measurable learning improvements at a cost that South Africa’s schools can sustain? Its future will depend less on the novelty of having a robot in class and more on the evidence of what that robot actually helps learners achieve.
