A “digital harness” is urgently needed to prevent artificial intelligence from outrunning human control as the technology approaches the “singularity”, global leaders were told today.
In comments released during the World Economic Forum Annual Meeting in Davos, the head of Japanese technology giant NEC warned that the moment when machine capabilities surpass human intelligence is “becoming more than just a theory”.
Takayuki Morita, CEO of NEC, argued that recent milestones — from reasoning models to autonomous agents — demonstrate that specialised AI systems can now outperform humans in complex tasks ranging from strategic decision-making to vaccine development.
“The lesson is that accelerating capability without ensuring control invites risk,” Morita said. “What is needed is not to slow innovation but to harness it.”
Wild horses
He drew a parallel to the invention of the physical harness thousands of years ago, which allowed humanity to utilise the “wild horse’s full horsepower” as a cornerstone of civilisation.
“Today, society must invent its digital equivalent – the systems and standards that ensure AI serves humanity responsibly,” he added.
Rather than relying solely on fragmented government policies, NEC advocated a “regulatory minimalism by design” approach, in which safeguards are engineered directly into the technology.
Morita outlined three specific technical innovations that would form this harness:
- Real-time hallucination detection: Systems that flag uncertain outputs and use “provenance watermarking” to distinguish synthetic content from verified sources, helping to strengthen trust in sectors like journalism and healthcare.
- Biometric-blockchain fusion: Anchoring digital identity in cryptographically verifiable records to track consent and prevent fraud, though Morita acknowledged this raises complex privacy trade-offs.
- Localised deployments: Keeping sensitive data within national or organisational boundaries to protect against geopolitical shocks and ensure data sovereignty.
The chief executive cautioned that without these mature technical safeguards, governments are turning instinctively to prescriptive regulations that risk splitting the global innovation landscape.
“If every jurisdiction creates divergent policy frameworks, AI development may concentrate in a few regions, limiting equitable access to its benefits,” Morita warned.
He argued that a principles-based approach would allow innovation to flourish while ensuring that “trust is engineered into systems from the outset”.
“To harness AI wisely is to build a world in which innovation creates shared prosperity, not just efficiency; trust, not dependence; and dignity, not displacement,” he concluded.
