The $4.8 trillion AI trust crisis: Why public-private partnerships are key for equitable innovation

Artificial intelligence (AI) adoption in enterprises surged an unprecedented 115% from 2023 to 2024, yet trust remains precarious – only 62% of business leaders and 52% of employees believe AI is deployed responsibly within their organizations.

This growing trust deficit urgently demands coordinated solutions that no single entity can provide like public-private partnerships (PPPs) can. PPPs combine government legitimacy, industry capability and civic oversight to turn “trust” into measurable controls, audits and redress.

The arithmetic is unforgiving: without trustworthy AI governance, the global economy forfeits not merely growth, but an estimated $4.8 trillion in unrealized economic upside by 2033 – largely the value lost to a widening digital divide between countries and communities with AI access and those without.

Across sectors, distrust has visibly hindered the development of AI. Even within companies, a KPMG study found only 35% of decision-makers trusted AI and analytics in their own operations. This is precisely where PPPs co-design standards, transparency, audit and accountability, so adoption is “safe by default”.

These trust gaps translate to missed opportunities: projects are shelved, efficiencies are unrealized and innovations are left on the table. According to a recent MIT study, 95% of AI pilots fail – far below expectations – due to model output inaccuracy, security concerns and ineffective change management.

World Trade Organization research shows that universal AI adoption could boost global trade by an additional 14 percentage points by 2040 – double what's possible under current fragmented approaches. In simpler terms, everyone benefits more when AI tools are trusted and used worldwide.

However, International Monetary Fund (IMF) analysis reveals a troubling pattern: currently, wealthy nations capture twice the productivity benefits from AI compared to developing economies, widening rather than narrowing global economic disparities.

Targeted PPPs – combining multilateral financing, national regulators and industry consortia – are the fastest mechanism to diffuse capabilities, data access and skills to the Global South, narrowing the productivity gap the IMF highlights.

The very reasons there is a trust deficit in AI – i.e. concerns about bias, privacy, security, safety, and accountability – span technical, social and regulatory domains. PPPs connect public guardrails with private innovation and bring academia/nongovernmental organizations to stress-test real-world impact.

When these stakeholders work in concert, their combined credibility creates systems worthy of confidence. For example, Estonia’s 99% online tax filing and EU-leading collection efficiency show how trust unlocks digital uptake.

Three-quarters (75%) of CEOs recognize that trusted AI requires effective governance, yet only 39% report having adequate frameworks in place. Leaders need a repeatable stack that turns policy into controls, controls into evidence and evidence into incentives.

The TrustworthyAI Index's assessment methodology – which builds on established frameworks including the Organization for Economic Co-operation and Development (OECD) AI Principles, Stanford Human-Centered Artificial Intelligence (HAI) benchmarks and National Institute of Standards and Technology (NIST) Risk Management Framework (RMF) standards – identifies that only 20% of leading models meet exemplary transparency and accountability thresholds.

Governments are catching up too: only 30% of countries have used AI in policy-making, signalling missing linkages between principles and line-of-business execution. This gap is structural, not attitudinal. As Microsoft CEO Satya Nadella noted at the World Economic Forum Annual Meeting 2024 in Davos: “I don’t think the world will put up anymore with something not thought through on safety, equity and trust.”

PPPs operationalize trust by turning principles into sector protocols, building assurance through certification/audits/incident reporting, and enabling responsible data-sharing for robust, fair models. PPPs deliver through three levers – with clear owners and artefacts:

The Partnership on AI (PAI), for example, convenes 129 technology companies, media organizations and civil society to establish concrete AI governance frameworks. Its Responsible Practices for Synthetic Media set provenance norms and are supported by organizations, including Adobe, BBC, OpenAI, TikTok and WITNESS.

Meanwhile, PAI’s Safe Foundation Model Deployment guidance and the AI Incident Database provide shared risk infrastructure, complementing PPPs by turning principles into verifiable practice.

The $4.8 trillion upside materializes only when trust closes the digital divide. As United Nations Secretary-General António Guterres argued at the UN Security Council, governments must urgently collaborate with technology companies on risk management frameworks, while systematically expanding access to ensure developing economies capture AI's transformative potential.

The Group of Twenty (G20) has already called for interoperable governance, safety assurance, and inclusive digital infrastructure. PPPs are the delivery vehicle – aligning standards, audits and deployment across borders.

Within 12 months, every G20 economy should:

Following that AI roadmap advances Sustainable Development Goals SDG 9 (Industry, Innovation and Infrastructure) and SDG 16 (Peace, Justice and Strong Institutions) and turns principles into growth – fast and fair. The opportunity is $4.8 trillion; the path is clear; execution must now be collective.