The Global Digital Compact Has the Right Vocabulary. It Still Needs Teeth.

The Global Digital Compact has the vocabulary of the future: inclusive digital economy, human rights online, digital public goods, data governance, AI cooperation, closing digital divides, trustworthy technology, capacity building, and safe digital spaces.

The language is broadly right. That is not nothing.

For years, digital policy has been split between three unsatisfying stories. Silicon Valley told a story of innovation as destiny. Authoritarian states told a story of sovereignty as control. Liberal democracies told a story of rights and markets, but often exported platforms faster than accountability. The Global Digital Compact tries to create a shared diplomatic frame for digital cooperation. It recognises that digital governance is no longer a niche technical issue. It is a development issue, rights issue, security issue, economic issue, and democratic issue.

But global digital governance has a familiar problem: principles are easier to sign than to enforce.

The Compact’s promise depends on whether it can avoid becoming another high-level document that says the right things while leaving the hardest questions to underpowered institutions and overburdened civil society.

There are four accountability gaps.

First, the implementation gap. Many states can endorse human rights online while passing vague cybercrime laws, expanding surveillance, blocking platforms, or ordering internet shutdowns. A compact that does not create meaningful reporting, monitoring, and consequences risks becoming diplomatic cover.

Second, the corporate power gap. Digital infrastructure is heavily shaped by private companies: cloud providers, app stores, platforms, AI labs, telecom firms, payment systems, and data brokers. Global governance cannot succeed if companies are treated as helpful stakeholders rather than powerful actors with human rights responsibilities. The UN Guiding Principles on Business and Human Rights already provide a baseline: companies must respect human rights and provide remedy where they cause or contribute to harm.

Third, the capacity gap. Many Global South governments and civil society organisations are expected to participate in AI and digital governance debates without comparable technical resources, research access, legal capacity, or bargaining power. The World Bank’s Digital Dividends report made a related point years ago: digital technologies alone do not produce inclusive benefits without complementary investments in institutions, skills, and governance.

Fourth, the evidence gap. Global digital policy often names harms without building systems to measure them. Internet shutdowns, algorithmic discrimination, platform-enabled harassment, AI incidents, data misuse, and online gender-based violence need documentation infrastructure. Without evidence systems, accountability depends on whoever can shout loudest or lobby hardest.

The Compact should therefore be treated as a beginning, not a settlement.

A stronger version of global digital cooperation would include several practical commitments.

One, states should report annually on internet shutdowns, platform blocking, surveillance procurement, cybercrime-law enforcement, and takedown requests. These reports should be independently reviewable.

Two, companies should publish human rights impact assessments for high-risk markets, not just global principles. Ranking Digital Rights’ work shows why corporate transparency must be evaluated across governance, privacy, and freedom of expression—not accepted at face value.

Three, AI governance commitments should include incident reporting, auditability, researcher access, and remedy. The NIST AI RMF and OECD AI Principles provide useful structures, but implementation should be tied to documented risk management, not press-release ethics.

Four, civil society in the Global South should receive funding for technical evidence systems: monitoring tools, secure data infrastructure, litigation support, digital forensics capacity, and independent research access.

Five, digital public infrastructure should be assessed for rights risks before deployment. Inclusion without safeguards can become efficient exclusion.

The Global Digital Compact deserves credit for naming the right terrain. But the next stage must be institutional. Who monitors commitments? Who can challenge violations? Who funds independent scrutiny? Who has access to data? What happens when states or companies ignore the principles?

Global digital governance cannot survive on vocabulary alone.

The world has enough declarations saying technology should serve people. The harder task is building systems that make powerful actors prove it.