- Physical goods face mandatory safety certification; AI largely does not.
- SA businesses lack a regulatory baseline for AI vendor due diligence.
- Liability for AI errors currently falls on the deployer, not the builder.
The Regulatory Paradox
Microsoft President Brad Smith has drawn a striking comparison, noting that milk cartons and washing machines are subject to more rigorous safety regulations than artificial intelligence. This observation underscores a growing global disconnect: physical consumer products face established, enforceable safety frameworks, while AI systems—capable of widespread societal impact—operate in a largely self-governed space.
For South African businesses, this regulatory vacuum creates both uncertainty and risk. Without clear legislative guardrails, local SMEs adopting AI tools cannot rely on compliance certification to vet vendors, potentially exposing them to liability, bias, or security flaws that a regulated market would typically mitigate.
- Physical goods face mandatory safety certification; AI largely does not.
- SA businesses lack a regulatory baseline for AI vendor due diligence.
- Liability for AI errors currently falls on the deployer, not the builder.
Why the Gap Matters for Local Adoption
The absence of AI-specific regulation slows enterprise adoption in regulated sectors like finance, healthcare, and telecommunications. South African decision-makers often delay AI projects because they cannot demonstrate compliance with non-existent standards, putting local firms at a competitive disadvantage against peers in jurisdictions moving faster on policy, such as the EU.
Job seekers and skills providers also feel the impact. Training curricula struggle to incorporate 'compliant AI development' when compliance definitions are fluid. This delays the pipeline of work-ready talent that South African service providers need to export AI services globally.
- Regulated sectors (fintech, health) hesitate to deploy AI without legal clarity.
- Skills gap widens as training lacks a regulatory target.
- Export readiness for SA AI services stalls without alignment to global norms.
Global Momentum vs. Local Inaction
While Smith's comments target a global audience, the pressure is mounting on South African policymakers. The EU AI Act sets a de facto global benchmark; local firms serving European clients must comply regardless of domestic law. Simultaneously, the African Union's Continental AI Strategy signals intent for harmonised governance, suggesting national legislation will follow.
Ignoring this trajectory is risky. South African providers who build 'regulation-ready' architectures now—documenting model provenance, data lineage, and risk assessments—will avoid costly retrofits later. Those waiting for the Gazette may find themselves locked out of procurement pipelines that demand proof of responsible AI practices today.
- EU AI Act creates extraterritorial compliance pressure for SA exporters.
- AU Continental AI Strategy hints at upcoming harmonised rules.
- Early 'regulation-ready' design reduces future retrofit costs.
Practical Next Steps for Stakeholders
SMEs should immediately adopt a voluntary AI governance framework—such as ISO/IEC 42001 or the NIST AI Risk Management Framework—to structure internal controls. This provides a credible signal to clients and insurers while bridging the legislative gap. Providers must document model cards, data sheets, and incident response plans for every deployed system.
Job seekers and training institutions should prioritise 'Responsible AI' certifications and practical auditing skills. Policymakers and industry bodies must accelerate the draft national AI policy into enforceable regulation, explicitly defining risk tiers and conformity assessment pathways to give the market the certainty Smith argues is missing.
- Adopt ISO/IEC 42001 or NIST AI RMF voluntarily today.
- Mandate model cards and data sheets for all production AI.
- Upskill in AI auditing and responsible AI certification.
- Advocate for clear risk-tier definitions in upcoming SA policy.
This Connect analysis references reporting or documentation from TechCentral. We add our own South African business context rather than republishing the original article.
View original source78 currently score 70 or higher for confidence. logistics leads the seven-day service signal set. Procurement and tender sources are excluded.
Open today's Connect Brief