Putting AI to Work: Solving the Productivity Challenge through Faster Adoption

Abstract

The panel explored how public policy can accelerate the adoption of productive AI technologies. Amazon’s David Zapolsky highlighted the role of digital public infrastructure, cultural change within large enterprises, and the need for open, non‑restrictive policy environments. UK Minister Kanishka Narayan outlined a four‑pillared adoption strategy—people, proof, place, and public‑relations—illustrating it with the Barnsley “AI Growth Zone” pilot and a nationwide compute‑investment program. The discussion moved to digital sovereignty, with Zapolsky defending open‑border cloud models while noting AWS tools for sovereign‑cloud capabilities. Shubhii Agarwal stressed early AI education, up‑skilling of the workforce, and creating safe‑to‑fail environments for employees. The session closed with a rapid round‑robin on how Global‑South nations can leapfrog into AI‑driven productivity.

Detailed Summary

The moderator thanked the audience repeatedly (a transcription artefact) before handing the floor to David Zapolsky. Zapolsky framed the conversation around two intertwined themes:

• Digital Public Infrastructure (DPI) – Cloud‑based services that the Indian government provides to citizens, which can be a “huge needle‑mover” when AI is layered on top.
• Policy Environment – The need for regulations that enable innovation rather than create “policy backwaters” that hinder cross‑border technology migration.

He added that cultural change within large enterprises, including Amazon’s own legal and public‑policy groups, is essential: employees must shift from “this is cool, I won’t look at it now” to actively exploring AI tools that can boost productivity.

“Once you do that, you start seeing transformation, you start seeing improved productivity.”

Zapolsky also praised the Indian government for opening the summit, noting that such gestures help move a culture from acceptance to thirst for AI use‑cases.

2. UK Perspective – Adoption as a Strategic Policy Goal

The moderator introduced Kanishka Narayan, Minister for AI, and invited a high‑level view on AI adoption. Narayan’s response was organized around four pillars he labelled People, Proof, Place, and PR (public‑relations).

2.1 People

• Adoption spreads fastest when pioneers quietly champion AI and share success stories.
• The UK government is cultivating AI ambassadors and sector champions to “bang the drum” for AI benefits.

2.2 Proof

• Systemic adoption requires robust evidence of ROI in both public and private sectors.
• The UK is scaling pilot projects into fully evaluated deployments, thereby building a data‑driven case for broader rollout.

2.3 Place

• To avoid an “London‑only” AI boost, the government has chosen Barnsley as a demonstrator town.
• In Barnsley:
  • Healthcare – AI shortens patient waiting times.
  • Education – AI narrows learning gaps.
  • SMEs – Local businesses pilot AI for competitive advantage.

2.4 PR

• Effective communication of both opportunities and nuanced risks is essential for public buy‑in.

Narayan then expanded on how the UK can diffuse adoption geographically:

• AI Growth Zone Programme – A massive compute‑capacity investment across the UK:

Region	Investment (bn $)	Jobs (≈)
South Wales	10	5,000
North Wales	10	5,000
Lanarkshire, Scotland	8	3,500
North‑East England	10‑15	—

• The compute hubs will be shared with schools (AI experiments), hospitals (clinical workloads), and local startup ecosystems (energy, finance, semiconductors).
• By geographically tying compute resources to local talent pipelines and industry clusters, the policy aims to create durable, community‑level AI ecosystems.

3. Digital Sovereignty – Tension Between Open Borders and National Control

The moderator steered the conversation back to David Zapolsky to address the rising global discourse on digital sovereignty.

• Zapolsky argued that restricting data flows generally inhibits innovation because Amazon’s business models rely on open borders for cloud, media, and retail services.
• He emphasized the long‑term nature of tech investment: Amazon’s multi‑billion‑dollar commitments in India (US $40bntodate,plusaplannedUS$35 bn) are intended to outlast political cycles.

3.1 What Sovereignty Means in Practice

• Customers need control over data residency, encryption, and access – not necessarily a national ownership of the cloud provider.

• AWS offers a Digital Sovereignty pledge allowing:

  • Choice of storage region(s)
  • Data‑at‑rest and in‑transit encryption with customer‑managed keys
  • “Sovereign cloud” configurations where only the customer can decrypt data, even if a government orders access.

• Zapolsky concluded that technology partners (including AWS) can meet sovereign‑cloud requirements while preserving the openness needed for rapid AI adoption.

4. Skills, Education & Workforce Enablement

The moderator turned to Shubhii Agarwal (LocoBuzz) for the human‑capital angle.

4.1 Early AI Literacy

• Curriculum change is required: AI skills must be embedded in school programs so students grow comfortable “talking” to AI models.
• Early exposure turns AI from a “homework‑doing machine” into a productivity partner.

4.2 Corporate Upskilling

• LocoBuzz’s hiring pipeline shows a 3‑6 month lag before new hires become proficient.
• This lag can be shortened if candidates already possess AI fluency.

4.3 Cultural Adoption & Safe‑to‑Fail Environments

• Initial rollout of cloud tools at LocoBuzz faced reluctance; success required a cultural shift where “mistakes are tolerated.”
• The company now provides AI sandboxes for both engineering and non‑engineering staff to experiment, learn, and iterate without fear of punitive repercussions.

4.4 Broader Policy Implications

• The central policy question is whether AI extends or subverts human agency.
• To extend agency, governments must ensure:
  • Foundational cognitive skills (distinguishing good vs. bad information, logical chain‑of‑thought) in education.
  • Workforce reskilling so employees can supervise AI agents, detect erroneous outputs, and intervene appropriately.

5. Round‑Robin: Opportunities for the Global South

Each panelist answered a rapid prompt: What is the biggest attainable AI‑adoption opportunity for Global‑South nations?

Speaker	Key Point
Kanishka Narayan	The AI for Development program focuses on small, efficient language models tailored to local workflows, and on open‑source infrastructure (AI incubator, I‑Security Institute) that can be shared globally.
David Zapolsky	AI can be a leapfrog technology—just as mobile phones bypassed PCs, AI can fast‑track development. Nations should be cautious about importing premature regulatory frameworks and instead leverage open‑border cloud services to accelerate growth.
Shubhii Agarwal	The young, tech‑savvy population plus scalable infrastructure and skills at LocoBuzz position India (and similar economies) to lead. Incremental, “slow‑and‑steady” adoption with strong skill‑building will eventually tip the balance.

6. Closing Remarks

The moderator thanked the panel, highlighted the session’s central thesis—that AI adoption is a distinct policy challenge requiring coordinated action across individuals, organizations, and institutions—and signaled the end of the discussion.

Key Takeaways

• Policy must be an enabler, not a barrier. Open, interoperable regulatory frameworks are essential for both cross‑border cloud services and domestic AI diffusion.
• Four‑pillar UK strategy (People, Proof, Place, PR) offers a replicable blueprint: nurture ambassadors, demonstrate measurable ROI, pilot in a defined locality, and communicate successes transparently.
• Digital sovereignty can coexist with openness when cloud providers give customers granular control over data residency, encryption, and key management.
• Early AI education and safe‑to‑fail corporate cultures are critical to shorten the skills gap and build confidence among workers.
• Geographically distributed compute investments (UK AI Growth Zones) tie hardware resources to local talent pipelines, fostering regional AI ecosystems beyond traditional tech hubs.
• Global‑South leapfrogging: focus on lightweight, locally‑adapted language models and open‑source tooling; avoid importing rigid regulatory regimes that could stifle innovation.
• Cross‑sector collaboration (government, big tech, startups) is already yielding concrete pilots (e.g., Barnsley, LocoBuzz’s internal sandboxes) that demonstrate measurable productivity gains.
• Adoption speed vs. institutional speed mismatch remains a challenge; sustained dialogue (as exemplified by this panel) is needed to align fast‑moving technology with slower policy cycles.

---

Prepared from the verbatim transcript of the “Putting AI to Work” panel at the AI conference in Delhi.

See Also:

• implementing-ai-standards-for-global-prosperity-in-an-era-of-agentic-ai
• thriving-with-ai-human-potential-skills-and-opportunity
• aligning-ai-governance-across-the-technology-stack
• responsible-ai-in-action-how-global-enterprises-are-building-trust-at-scale