The ‘Brussels Effect’ Blueprint: We’re Importing the EU’s AI Act While Drafting a Three-Way Regulatory War
This past week, the European Commission, in an act of pragmatic humility, tabled its “Digital Omnibus proposal.” In plain language, Brussels is already patching its new digital laws, including the landmark AI Act. Why? Because it discovered that rigid regulation creates “heavy administrative burdens” and strangles the SMEs, it needs to compete.
For us in Ghana, this is not a distant academic exercise. It is a direct and urgent warning.
Our own proposed Emerging Technologies Bill (ETB) is a clear import of the EU’s core philosophy. We are adopting the “Brussels Effect”—a smart, strategic choice to align our laws with the world’s most comprehensive regulatory standard.
But this strategy is failing in its execution. We are making a dual error:
- We are at risk of importing the very friction (the “heavy administrative burdens”) that the EU is now desperately trying to fix.
- More dangerously, we are actively and simultaneously drafting multiple, conflicting laws that will place our entire tech ecosystem in a state of legal paralysis.
As the Ministry of Communication, Digital Technology and Innovations (MoCDTI) prepares these bills for Cabinet approval, and as Parliament readies itself for consideration, we must avert this looming legislative collision.
1. The ‘Start-Up Paradox’ (The Friction We Are Importing)
We currently have two flagship bills moving in opposite directions.
On one hand, the Innovation & Start-Up Bill (ISB) is a bill designed to accelerate, offering incentives and simplified registration to unleash our entrepreneurs.
On the other hand, the Emerging Technologies Bill (ETB), by importing the EU’s high-risk, rights-based architecture, is a heavy anchor. It seeks to impose a sweeping compliance framework on all new tech.
Here is the fundamental contradiction: What good is a tax holiday if a start-up is bankrupted by compliance costs before it ever earns a cedi?
The EU’s new “Omnibus” proposal is scrambling to add “regulatory privileges for SMEs.” Our legislation must start there.
2. The ‘Bias Paradox’ (The Critical Component We Missed)
This is the most urgent technical flaw. Both our National AI Strategy and the draft ETB rightfully demand, just as the EU Act does, that AI systems be “transparent, bias-free, and ethical.”
But how do you prove an AI is not biased?
You must test it.
To test a loan app for ethnic or gender bias, a developer must process sensitive data on ethnicity and gender. Under a strict reading of our own Data Protection Act, 2012 (Act 843), doing so is legally treacherous.
The EU’s lawyers saw this coming and engineered a solution directly into the original, final text of the AI Act. That solution is Article 10(5). This provision creates a narrow, specific, and vital “Bias Safe Harbor,” giving developers a legal basis to process sensitive data under strict safeguards, for the sole purpose of detecting and correcting bias.
Our draft appears to have imported the requirement (“be bias-free”) while forgetting to import the legal mechanism (“Article 10(5)”) that makes it possible. As written, our ETB is in direct, unavoidable legal conflict with our Data Protection Act.
3. The ‘Legislative Collision’ (The Three-Way War We Are Drafting)
The EU’s new simplification package is focused on clarifying the “interplay” between its many digital laws. We are not just ignoring this lesson; we are doing the exact opposite.
We do not need to imagine a future conflict. We are actively drafting a three-way jurisdictional war right now.
Bill 1: The Emerging Technologies Bill proposes a new Emerging Technologies Agency to govern the ethics and deployment of AI.
Bill 2: The Cybersecurity (Amendment) Bill proposes giving the Cyber Security Authority (CSA) new power to certify the security of AI products.
Bill 3: The Data Protection (Amendment) Bill seeks to modernize Act 843, giving the Data Protection Commission (DPC) stronger authority over the data that fuels AI.
Crucially, this Data Protection amendment does not solve the “Bias Paradox.” It is a siloed update that fails to build the legal bridge to the ETB.
This is not a regulatory framework; it is regulatory cholesterol. It is a three-way battle over who governs AI, and it will choke innovation and create crippling uncertainty for investors before a single law is passed.
4. A Smarter Strategy: A Call to the Ministry, Cabinet, and Parliament
We do not need to reinvent the wheel. We simply need to be smarter importers and coherent legislators.
This requires a clear-eyed intervention at every stage.
- The MoCDTI must de-conflict these bills, integrating these 2.0 fixes before they are finalized for Cabinet.
- Cabinet must scrutinize these bills for their interoperability as a primary condition of approval.
- Parliament, specifically the relevant Select Committees, must serve as the final backstop, refusing to pass one bill without knowing how it interacts with the other two.
The specific demands on these bodies are clear:
- Define a Clear Hierarchy NOW: The Ministry must force these three bills to speak to each other. A “Statement of Interoperability” is essential. Which agency is the lead on AI? Which is in support? Who has the final say on certification versus ethics?
- Legislate the “Bias Safe Harbor” IMMEDIATELY: The ETB must have its own “Article 10(5).” It must create a clear, legal safe harbor for processing sensitive data only for auditing and correcting bias.
- Mandate Proportionality: The ETB must be explicitly aligned with the ISB. This means importing the SME privileges the EU is now adding, giving our startups a simplified compliance track from day one.
A blueprint that forgets its own legal engine is a recipe for failure. A government that passes three conflicting laws at the same time is not regulating; it is creating chaos.
By: Sitsofe Mensah, a technology policy enthusiast and a writer for the IMANI Centre for Science, Technology and Innovation Policy (CSTI)
