AI Trust: A Global Standards Solution

The digital landscape, once a realm of burgeoning possibility, now teeters on the precipice of a trust crisis.

Artificial intelligence, a technology hailed as transformative, has, in its rapid and unchecked proliferation, become a double-edged sword.

It is no longer just about sophisticated algorithms performing complex tasks; it’s about a pervasive, cheap, and increasingly indistinguishable stream of manufactured reality.

From eerily convincing deepfakes that can tarnish reputations or sow discord, to subtly biased content and outright fabrications, the very bedrock of verifiable information is eroding.

The danger is clear: a future where the authenticity of any image, video, or piece of text, regardless of its origin, is cast into doubt.

Yet, amidst this growing chaos, a quiet revolution is underway, spearheaded by international bodies often associated with the mundane, yet essential, architecture of modern life.

At the recent UN “AI for Good” Summit in Geneva, Switzerland, the gravity of this challenge was palpable.

Philippe Metzger, secretary-general and CEO of the International Electrotechnical Commission (IEC), a body historically focused on the standards that power our homes and industries, articulated a vision for reclaiming trust.

His organization is not alone; it’s part of a formidable alliance, the AI and Multimedia Authenticity Standards Collaboration (AMAS), joining forces with giants like the International Organization for Standardization (ISO) and the International Telecommunication Union (ITU).

Their collective mission: to rein in AI’s chaotic sprawl, one painstakingly crafted standard at a time.

This endeavor, as Metzger himself concedes, is akin to “herding cats.”

The AI market is a swirling vortex of innovation, diverse players, and technologies that evolve at breakneck speed.

To impose order on such a dynamic ecosystem seems a Sisyphean task.

But the AMAS coalition believes standards—specifically those focusing on watermarking, digital verification, and traceability—offer a critical pathway to discern digital fact from fiction.

The categories of standards under development within AMAS are comprehensive, touching on critical aspects like content provenance, trust and authenticity, asset identifiers, and even human rights declarations.

The groundwork has already been laid; a foundational standard for AI trustworthiness, providing guidelines for assessing system reliability, was published as early as 2020.

More recently, the IEC and ISO unveiled the first part of their JPEG Trust series, a crucial weapon against the rising tide of deepfakes in visual and audio media.

This standard zeroes in on provenance, detection, fact-checking, and content credentials for JPEG images, marking a tangible step towards verifiable digital content.

Future standards are slated to cover digital watermarking, general trust guidelines, and a robust framework for authenticating multimedia content.

A core tenet of these international standards bodies is that their guidelines are, by design, voluntary.

“None of these standards are mandatory, and that’s not going to change.

That’s core DNA of what we do and stand for,” Metzger affirmed.

This immediately raises the million-dollar question: how can voluntary guidelines hope to tame a technology that respects no boundaries?

The answer, according to Metzger, lies in a multi-pronged approach driven by market forces, formal assessments, and, potentially, government mandates.

The power of market pressure cannot be overstated.

In an increasingly interconnected digital economy, companies that can demonstrate adherence to widely accepted standards will gain a significant competitive edge.

Such conformity helps them integrate seamlessly into broader ecosystems and, crucially, builds consumer confidence.

“To have fair dealings, fair trade, and fair exchanges, you need a reference point,” Metzger explained.

This reference point, provided by a common standard, becomes a de facto requirement for doing business responsibly.

Beyond market dynamics, formal conformity assessments are emerging as a powerful enforcement mechanism.

Metzger highlighted IEC’s existing service for assessing claims around quality and carbon footprints, noting how companies making bold pronouncements about their “super green” products often seek third-party verification.

This model, he anticipates, will be applied to AI quality and trustworthiness standards.

An audit, a stamp of approval from an independent body, could become the gold standard for companies seeking to prove their AI systems are not just innovative, but also reliable and ethical.

While governments might not directly mandate the standards themselves, they could certainly mandate that companies using AI undergo such third-party verification, effectively creating a regulatory pull for voluntary standards.

The most formidable challenge, however, remains the dizzying pace of AI innovation.

Standards bodies, by their very nature, are consensus-driven organizations.

Their processes are deliberate, designed to ensure broad agreement and robust guidelines.

This methodical approach stands in stark contrast to an AI landscape where entirely new paradigms, such as agentic AI, can emerge and dominate development in less than a year.

“The speed is really a challenge,” Metzger conceded, acknowledging the difficulty in keeping pace with rapidly multiplying dimensions of AI.

While IEC committees are already scrambling to develop standards for agentic AI, the inherent slowness of the process means fully baked standards might always lag behind the cutting edge.

Yet, Metzger maintains a characteristic optimism.

He acknowledges the rapid currents of scientific and market drivers but points to an inherent “auto-corrective mechanism built into society, including governance systems.”

This suggests that even without direct regulation, the societal pushback against unchecked AI—from privacy concerns to misinformation—will eventually create resistance at various levels, through different channels, and from diverse stakeholders.

This collective societal push, he believes, will ultimately compel the industry towards greater responsibility and, by extension, adherence to the very standards being painstakingly crafted today.

The battle for trust in the digital age is far from over, but the architects of authenticity are laying down their blueprints, hoping to build a more verifiable, and therefore more trustworthy, future.