From Hype to Habit: Building Institutional Trust in AI

Session
Partners

Session Summary

Important
Quotations

“We can only do that with AI and with digital platforms because there’s not enough nurses, teachers, hospitals available.”
Paschal Donohoe
“Trust is all about doing the right things when nobody’s looking… it’s gained in drops and lost in buckets.”
Hossein Nowbar

Key
Takeaways

  • Nowbar said the technology is ready and the real gap is in deployment and trust, citing ServiceNow’s AI Maturity Index of 4,500 executives, which found that the share experimenting with AI rose from 29% to 58% in a year, but only 9% have actually put AI into production. 
  • That trust deficit comes from a lack of governance, such as process mapping, audit trails, and feedback loops, which matters even more as AI agents act on people’s behalf using their credentials, the problem ServiceNow’s “AI control tower” is designed to address through data, intelligence, workflows, and security and governance. 
  • ServiceNow sees AI as augmenting human intelligence rather than replacing judgment, and its ServiceNow University has retrained 2 million of its goal of 3 million people by 2027 to build AI literacy. 
  • Donohoe said the first step for developing economies is infrastructure, since internet access exceeds 90% in many parts of the world but is below 30% in many World Bank partner countries, and the World Bank is exploring an “AI in a box” model for places with major infrastructure gaps. 
  • The World Bank’s World Development Report found that about 4–4.5% of jobs face labor market disruption from AI, while 16%, four times as many, could see significant productivity gains, which is why much of the developing world sees AI more as an opportunity than a risk. 
  • AI is already delivering results, such as customized literacy plans for every child in large Ghanaian classrooms and handheld maternity guidance for nurses in African primary care centers, though not yet at the scale needed, and Donohoe said goals like reaching 1.5 billion people with better health care can only be met with AI because there aren’t enough nurses, teachers, and hospitals. 
  • Nowbar described trust as a social contract that is “gained in drops and lost in buckets,” noting that society has historically caught up with new technologies through privacy, antitrust, and telecom regulation, so tech companies must balance innovation with accountability. 

Action
Items

  • Organizations should put governance in place before scaling AI, including audit trails, workflow monitoring, and controls over AI agents that act with user credentials. 
  • Companies should keep human judgment at the center of AI use and invest in reskilling and AI literacy for their workforce. 
  • Governments, development banks, and the private sector should invest in digital infrastructure and connectivity in developing economies, including low-infrastructure models like “AI in a box.” 
  • Developing countries and their partners should scale up proven uses of AI in public services such as education and health care. 
  • Policymakers in developing economies should focus on capturing AI’s productivity gains for growth, not only on managing its risks. 
  • Technology companies should adopt strong accountability and responsibility practices proactively rather than waiting for regulation to catch up.