Video summary

GCI World 2026 April Session14 Opening and Panel Discussion

Main summary

Key takeaways

News and Commentary

Summary of the Session (GCI World 2026 – April Session 14 Opening & Panel)

The event focused on how countries outside the main AI hubs (US/China)—specifically Bangladesh and Japan—can turn local AI talent into real economic growth, rather than simply consuming AI technologies.

A central theme was that AI progress must be distributed and human-centric, and that the future should not be shaped only by a small number of countries or large companies.

Opening Remarks (JAICA Bangladesh)

  • AI is framed as a foundational technology impacting nearly every sector: productivity, education, healthcare, agriculture, manufacturing, and public services.

  • JAICA emphasized that the question is no longer whether countries will use AI, but how they will create value from it.

  • Human-centric AI was highlighted as essential for: improving lives, expanding opportunities, and supporting inclusive, sustainable development.

  • Bangladesh’s opportunity is tied to its:

    • young population
    • strong engineering potential
    • ambition in digital development
  • JAICA’s goal goes beyond training individuals: it aims to cultivate people who can create businesses, solve local problems, generate employment, and contribute to national growth.

Panel Discussion: “Empowering local AI talent to accelerate economic growth—How can third countries catch the AI wave?”

Panelists discussed how AI contributes across education, startups, and industry cooperation, with growth depending on local problem-solving and scalable entrepreneurship.

1) AI education should shift from “coding alone” to hypothesis-driven work

One panelist (Aki / Matsulab global strategy office) argued that AI is changing education rapidly: students can use AI tools to move faster.

As “basic coding” becomes less central, the differentiators become:

  • hypothesis-driven thinking
  • rapid testing and iteration
  • applying AI in one’s local language/context (reducing previous barriers)

Education should enable students to build solutions that produce impact, not just learn models.

2) Startups in Bangladesh: use AI to remove repetitive work and build real business workflows

Shaher (founder of the Bangla-focused learning startup Shiko / Shaher) described AI as a “game-changer” for learning—especially personalized, democratized education for Bangladesh’s large student and youth populations.

On building real value, the startup’s approach was practical:

  • use AI tools (e.g., assistants/chat systems) to eliminate repetitive, data-heavy manual tasks
  • convert saved time into deeper work, iteration, and product/business execution
  • compete not by copying Silicon Valley directly, but by using access to tools and applying them aggressively every day

3) AI talent should “stay and build” locally via locally grounded, globally scalable problems

Shuai (founder building AI infrastructure) emphasized:

  • selecting vertical/domain-specific use cases rooted in local needs
  • using open models to solve problems global superpowers may not prioritize
  • aiming for globally scalable solutions even if the initial problem is local

Japan was also referenced as facing early pressure from labor shortages and aging, creating opportunities for AI solutions sooner than in other places.

4) International industry cooperation must evolve from “outsourcing hours” to “solution delivery”

A Japanese industry leader at BGIT Limited (Dan) explained that client expectations have shifted:

  • customers want business solutions, not just cheap labor or development hours

BGIT’s response is to reposition as an AI solution provider, helping engineers shift mindset from:

  • “deliver code” → solve business challenges

The panel also noted that clients may not always know exactly what they need, so strong technical partners add value by guiding solution design and modernization.

5) Physical AI (robotics/control) is seen as a future opportunity for third countries

Renzo Kasher (Lenzosan) argued that general-purpose AI tools don’t directly solve every country’s specific problems. Generalized models often struggle in high-context, real-world settings, which opens entry points for countries like Bangladesh earlier in frontier domains such as robotics and physical AI.

Examples included specialized physical AI needs in:

  • renewable-energy systems
  • large pilgrim logistics
  • dense cities and tourism management

These environments often require specialized physical AI approaches, not just “chat-based” AI.

“Survival Strategy” for Third Countries: Not Defense—Be Problem-First and Hypothesis-Driven

In open discussion, multiple speakers aligned on a similar strategy for Bangladesh/Japan and other emerging economies:

  • Don’t treat the goal as competing to build the largest frontier models
  • Instead focus on:
    • context-specific problems
    • building on top of available infrastructure (tools/LLMs) with localization and fine-tuning
    • developing thesis/hypotheses about the future of a domain, since standards are still unsettled

The recurring message: future leaders will be those who iterate quickly using AI tools, spending saved time on creative hypothesis-building and solution design, not only prompt-based execution.

There was also strong emphasis on foundational understanding (especially mathematics/physics) to use AI effectively and evaluate limitations.

Audience Q&A Highlights

  • Is computer science alone enough, or do students need deep domain knowledge (e.g., agriculture/business)? Panelists responded that AI doesn’t remove the value of domain understanding—problem-solving still requires context. Continued learning and strong foundations were encouraged.

  • How do you find “a problem worth solving with AI”? A startup founder advised:

    • start with the problem first (not “AI first”)
    • repeatedly challenge initial assumptions
    • accept that most ideas will be wrong—refining until the real “1%” worth solving becomes clear

Presenters / Contributors

  • Tekashi Wabuchi (Moderator, Masol University Tokyo)
  • Yuko Morikawa (Senior Representative, JICA Bangladesh Office)
  • Aki (Matsu Lab; researcher and Global Strategy Office)
  • Dan (Chief Revenue Officer, BJIT Limited / Bangladesh Japan IT)
  • Shaher (Founder, Shiko)
  • Shuai (Founder, “Baktan” / Matsu Lab startup building AI infrastructure for enterprise HR; remote)
  • Renzo Kasher (Lenzosan) (Researcher, Matsu Lab; Physical AI/robot national models)
  • Professor Maxu (mentioned for a message segment, but not heard directly in the provided subtitles)
  • Brock University representative(s) (closing remarks mentioned; no specific individual name captured in the subtitles)

Original video