Stay with the argument.

▶1:01

Practical method · DAY 1

Draw the decision before choosing the tool

Start with the workflow, human decisions and intended benefit.

▶1:18

Open question · DAY 1

Who funds safety after the pilot?

A practitioner challenges short grants that leave nobody to monitor deployed systems.

▶1:47

Research explanation · DAY 1

Match the evidence to the question

Why trials, qualitative knowledge and institutional analysis answer different questions.

▶2:23

Practical case · DAY 1

Interest is not the same as access

The adoption gap starts before installation: devices, data cost and household control.

▶3:11

Negative case · DAY 2

An ethical promise still needs user value

Fairbnb’s publicity did not solve marketplace density and matching problems.

▶2:57

Boundary case · DAY 2

Where cooperative ownership stops

Overseas subsidiaries expose the limits and conditions of a shared-ownership model.

▶1:02

Research explanation · DAY 2

Owning shares does not settle market power

A share of profits and influence over dominant firms are separate questions.

▶1:43

Practical case · DAY 2

Purpose protection and democracy are different decisions

Sharetribe’s founders retained voting control before a later democratic transition.

▶2:52

Research explanation · DAY 2

A seat at the table is not a veto

A study distinguishes representation, votes and the power to stop a decision.

▶2:09

Negative case · DAY 2

Why a shared registry closed

An API change exposed the recurring costs and legal responsibilities of shared data.

▶1:16

Boundary case · DAY 3

Traceability is only the beginning

Knowing origin does not settle due diligence, participation or the distribution of benefits.

▶1:54

Challenge · DAY 3

What should scarce minerals be used for?

The transition debate must also consider demand, sufficiency and rights holders.

▶2:36

Negative case · DAY 3

What if better audits lose customers?

More complete checks can reveal problems while creating a commercial penalty for the verifier.

▶3:23

Contested exchange · DAY 3

Who will actually pay farmers for data?

An objection is followed by proposals on bargaining power and audit savings—not evidence of payment.

▶0:40

Practical method · DAY 3

Reporting cannot build a recycling plant

The moderator explains why physical processing capacity is needed for reporting to lead to improvement.

▶3:33

Contested exchange · DAY 2

Possible alternatives, unequal ability to build them

Local-model optimism meets a challenge about resources and bargaining power.

▶1:14

Practical method · DAY 3

A risk model still needs human judgement

Methods, assumptions, purpose and escalation remain decisions people must own.

▶2:19

Challenge · DAY 1

Whose readiness are we measuring?

The final exchange brings data feminism, intersectionality and an Afrofeminist lens into the readiness discussion.

▶1:11

Practical perspective · DAY 2

AI infrastructure needs people

Nuria Oliver explains why AI infrastructure depends on talent, opportunity and support for people whose work is changing.

▶1:24

Panel synthesis · DAY 2

Markets, institutions and people

Bhaskar Chakravorti connects competition, institutions and public agency as starting points for changing who benefits from AI.

▶0:51

Financing model · DAY 2

Ownership beyond existing savings

Jens Lowitzsch explains how a renewable-energy ownership model uses trusteeship and future returns to include people without existing savings.

▶0:42

Practical case · DAY 2

Adopt the technology. Look after the people.

MONDRAGON’s Ander Etxeberria Otadui describes training, transfers and supported retirement as responses to automation.

▶2:36

Practical method · DAY 1

Research has to happen during deployment

Rachel Coldicutt explains why AI evaluation must continue during use, asking who is excluded when a tool fails and whether it works in each new context.

▶3:05

Practical proposal · DAY 1

Who will maintain AI when the project ends?

Rachel Coldicutt proposes shared maintenance and a culture of safety that continues beyond the pilot, the grant and the launch checklist.

▶1:32

Concrete case · DAY 1

Build AI for a context, not for everywhere

Drawing on research in southern India, Rachel Coldicutt describes small enterprises building for specific languages, needs and levels of access.

▶0:52

Core argument · DAY 1

Greater uncertainty makes evidence more necessary

Becky Faith argues that shrinking budgets and pressure to move quickly make research and equitable partnerships more important. She speaks in a personal capacity.

▶1:23

Practical case · DAY 3

Trust depends on the promises we keep

Victoria Snelling explains what Leather Working Group must preserve as it digitises: trusted expertise, responsible data handling and promises kept.

▶0:50

Practical case · DAY 3

Certification needs evidence people can see

Rob Shaw explains why spreadsheets and manual transfers make it harder to substantiate certification claims—and why saying a system is audited is no longer enough.

▶1:49

Concrete case · DAY 3

Forty shots on goal

Karan Chopra recalls a storm that destroyed a rice-processing facility in Ghana, then explains why a failed season can be far harder for a smallholder to survive.

▶1:02

Core argument · DAY 3

Why more data does not guarantee better decisions

Iggy Bassi argues that more data and AI tools do not automatically solve the decisions created by overlapping supply-chain risks.

▶1:47

Practical method · DAY 3

Verify the process behind the claim

Marc Boissonnet of the TIC Council explains a two-step approach: verify how sustainability data is produced, then verify the claims made from it.

▶1:00

Practical advice · DAY 3

Sustainability is business strategy

Lela Mélon urges businesses to use verified evidence to rethink their strategy and make sustainability work for their business.

▶1:10

Opening perspective · DAY 2

The three rungs of human agency

In his day-two opening, Martin Smith sets out three foundations of human agency: security to plan, a meaningful say in decisions, and a lasting share in the value people create.