51ÁÔÆæ

Ms. Leigh Anne DeWine,
Mr. Roberto Su¨¢rez Santos,
Ms. Whitney Baird,
Mr. Glenn Spencer,
Excellencies,
Distinguished guests,
Ladies and gentlemen,

Let me begin by warmly thanking Amazon for hosting us, and the International Organisation of Employers, the United States Council for International Business and the U.S. Chamber of Commerce for bringing this community together.

This morning, before most of us arrived, software somewhere decided who would be offered a shift, whose job application would reach a recruiter and whose performance would be flagged for review.

Most of the people affected did not see those decisions in the making. Many may never know that a machine helped make those decisions – or advance them.

And workers are not asking whether artificial intelligence is coming. They are asking whether it will work for them.

My theme for this session of the General Assembly is “Restoring Trust, Managing Transformation: A United Nations That Delivers for All.”

Nowhere do trust and transformation meet more directly than in the world of work.

Dear colleagues and friends,

The International Labour Organization (ILO) finds that one in four workers in the world is in an occupation with some exposure to generative AI. Most jobs are more likely to be transformed than eliminated.

That should give us confidence. But it should not give us comfort.

The effects will not be evenly shared. In high-income countries, women are almost three times as likely as men to hold jobs in the highest-exposure category.

Behind many AI systems is another workforce we rarely see: people who label data, moderate content and perform microtasks, often without adequate pay or social protection.

Some 2.1 billion workers, nearly 58 per cent of the global workforce, are in informal employment, beyond the reach of most labour protections.

And in July, the Independent International Scientific Panel on AI, established by the UN General Assembly, warned that without investment in skills, infrastructure and labour market institutions, AI risks widening inequality, displacing workers and shifting wealth from labour to capital.

None of this is inevitable.

AI is not the weather. It does not simply happen to us. It is designed, procured and deployed by people. And people and institutions must remain accountable for how it is used.

That is why this is a human rights imperative.

Distinguished participants,

Fifteen years ago, Professor John Ruggie, a colleague of mine, addressed this Forum. Soon afterwards, the Human Rights Council endorsed the Guiding Principles on Business and Human Rights that he had developed.

Those Principles predate generative AI, but they speak directly to this moment: States have a duty to protect; companies have a responsibility to respect; and people harmed must have access to remedy.

An algorithm changes none of that.

In 1944, the ILO declared that labour is not a commodity. In the age of AI, we must also remember: workers are not data points.

I come from a country where once workers paid the price when clear warnings went unheeded.

On 24 April 2013, Rana Plaza collapsed near Dhaka, Bangladesh, killing more than 1,100 people, most of them ready-made garment workers. Cracks had been reported. Yet workers were sent back inside.

The information existed. What was missing was the will to act on it – and the power of workers to say no.

What turned that tragedy into reform was shared responsibility.

Global brands and unions joined forces on factory safety. The Government, employers and workers agreed on a national plan.

The ILO supported both efforts, and thousands of factories were inspected and made safer. That work continues.

Today, AI can help companies detect the cracks in global supply chains: unsafe conditions, forced labour and other abuses. That is welcome. But visibility is not responsibility. Data must lead to action, and technology must strengthen – never silence – worker voice.

So what must we do? Let me offer three priorities.

First, companies.

Extend human rights due diligence across the full AI life cycle – from design and procurement to deployment and review. Before using a system to recruit, schedule, monitor, evaluate or dismiss workers, identify who could be harmed and act on the answer.

Look at both ends of the chain: those who train the systems, and those whose working lives are shaped by them.

Engage workers and their representatives through genuine social dialogue. Worker voice is not a cost. It is an early-warning system.

Invest in people before jobs change, not after.

Preserve meaningful human oversight over decisions that affect livelihoods. Give every worker a way to question a decision and a person who must answer.

AI can automate tasks. It must never automate accountability.

Second, Governments.

Business needs predictability. Workers need protection. Good practical rules can deliver both.

That requires labour laws that keep pace with changing forms of work; continuous social protection; and skills systems that reach every worker, including those in informal employment.

In June, the International Labour Conference adopted Convention No. 193 – the first international labour standard dedicated to the platform economy. It provides for transparency about automated systems, written explanations for significant adverse decisions, and review with appropriate human involvement.

This is practical multilateralism. It shows that social dialogue can produce safeguards while enabling innovation.

Third, international institutions, including the United Nations.

AI governance needs shared evidence and an inclusive table. The General Assembly has helped create both: an independent international scientific panel, and a Global Dialogue on AI Governance in which every country has a voice.

At the first session of the Dialogue in Geneva in July, participants called human rights due diligence and impact assessment central tools under the Guiding Principles, for States and companies alike.

The Dialogue meets next year in May, here in New York.

As President of the General Assembly, I will work to ensure that the world of work is heard at that Dialogue – workers and employers, young people, women, civil society, scientists and industry.

The future of work cannot be decided only by those who build the technology, or only by countries that can afford it.

This was the message from more than 600 young people who gathered at the UN in August to present a youth-led declaration on AI policy and governance.

All countries – including developing countries – must help shape that future and share in its benefits. We must close divides in infrastructure, data, skills and computing capacity. We must ensure that technological transformation serves people everywhere and leaves no one behind.

That is what I mean by “Innovation with Inclusion”.

Ladies and gentlemen,

Tomorrow morning, software will again influence who is offered a shift, whose application is seen, and whose performance is questioned.

Whether those decisions uphold human dignity is not the machine’s choice. It is ours.

Let us make AI accountable to the people whose work makes it possible.

Let us measure innovation by the decent work it creates, the inequality it reduces and the dignity it protects.

And let us remember: it is not enough to make workers fit for the future. We must also make the future fit for workers.

I thank you.