- Fewer than 1 in 100 organizations have fully adopted responsible AI practices.
- Delivering on the promise of AI requires trust, and current AI strategies are lacking in this area.
- The technology is poised to implement effective AI governance, but leaders must make it a priority.
Artificial intelligence (AI) has great potential, but its future depends on trust. And when it comes to making the most of this technology, the data speaks for itself. Less than 1% of organizations fully operationalize responsible AI practices. This gap is not just technological. It’s a structural thing.
Without governance built in from the start, AI risks repeating the mistakes of past technologies, from poor data quality to opaque decision-making and weak accountability. The World Economic Forum’s Driving Responsible AI Innovation: A Handbook report delves into what this means and how innovators and those using AI can realize its potential.
Governance is critical at the point where policy and product meet. When governance shows up late, it’s like pouring concrete after the residents have moved out. Today it’s a hairline crack, tomorrow it’s a structural problem. Incorporating it into your blueprints won’t slow you down. It stabilizes it, expands it, and sustains it.
Trust in AI starts at the data layer
The Advancing Responsible AI Innovation: A Handbook report highlights a simple truth: The success of modern AI relies on modern data governance. However, many organizations still suffer from siled systems, uneven data quality, and approval processes that slow progress and undermine trust.
Distributed ledger technology is starting to change this. EQTY Lab collaborates with NVIDIA to use “Verifiable Compute” to anchor encrypted receipts in Hedera. Tamper-proof records of how models are trained and inferred. ProveAI covers the other side, documenting who has been exposed to what training set, when, and under what policies, in line with new regulations like the EU AI law. It’s not a reactive responsibility, it’s a real-time responsibility.
These approaches demonstrate what happens when governance is built in from the beginning. Trust is not something added as an afterthought as a safety net. It becomes part of the system itself and is continuous, transparent, and resilient by design.
First, pour the foundation
It’s not just the data that matters. Organizations need owners. The World Economic Forum’s playbook calls for a gradual path, starting with designated AI stewards, cross-sector councils, and centralization, and maturing to federated oversight as capabilities grow. This avoids both confusion and bureaucracy.
Decentralized systems also offer useful lessons. In decentralized finance (DeFi), token holder voting and governance councils help balance speed, transparency, and resilience. Open source communities further strengthen accountability and distribute oversight across developers and users who audit code and protect its integrity. These models aren’t perfect, but they show that building governance into your design can make accountability a built-in strength, rather than an afterthought.
Taking this a step further, we can build governance into the system itself by creating a council of equally accountable businesses, nonprofits, and universities. No actor has unchecked power. This architecture has created enduring trust and responsible scale precisely because power is distributed rather than accumulated.
AI requires such discipline. Governance must be visible, intentional and continuous. Guide design, implementation, and growth. This builds resilience and strengthens trust.
A blend of progress and principles
Governments themselves need to clarify the AI value chain, especially as generative AI blurs the lines between creators, adopters, and users. Without clear accountability and shared standards, we invite systemic risk. International coordination is also important. Just as financial markets rely on common rules and oversight, AI will need guardrails across borders to inspire trust.
We have already reached the early stages. In the UK, the reintroduced AI Regulation Bill proposes an AI Authority and a mandatory AI Officer to oversee responsible deployment. The EU is taking a different approach, using the AI Act to force compliance across the bloc. These are specific examples of different models for addressing AI governance.
The current challenge is to improve these models. Define ownership, strengthen senior governance roles, embed oversight across deployments, and work toward global alignment. Trust, safety and innovation all depend on it.
Build AI on a solid foundation
Forum strategy calls this a decisive opportunity. AI can be a technology that people fear, or it can be a technology that people trust to advance progress while protecting their rights. The outcome will depend on whether governance is treated as a foundation or an afterthought.
Like any structure, once the foundation is established, anything built on top of it can stand taller and last longer. Innovation is more resilient and transparent when governance is built in from the beginning. Trust grows with adoption, giving us the opportunity to not only scale AI quickly, but responsibly, with accountability and inclusivity at our core.
Progress does not come from fenced-in initiatives. It requires an open ecosystem and serious collaboration between policy makers, builders and researchers. Make governance a catalyst, not a brake, on trust and growth.
