International cooperation needed to prevent AI widening income gaps, trade experts warn

“Concerted international cooperation” will be required to democratise the trade benefits of AI, a panel of experts speaking at the WTO public forum told delegates last week (13 September).

During a discussion on “AI and data-driven supply chains”, leading figures from the world of trade and development highlighted the immense benefits AI offers for increasing supply chain efficiency, accelerating the manufacture of complex products and troubleshooting common sources of disruption throughout distribution.

However, intervention to ensure a regulatory framework that will empower low- and medium-income countries to reap these benefits will be a necessity in the coming years to prevent AI advances widening existing digital and wealth divides, the speakers said.

‘Tipping point’

Microsoft’s senior director of UN affairs and international organisations, Ashutosh Chadha, was enthusiastic about the potential of AI to not only enhance supply chain efficiency but substantially democratise the benefits worldwide.

Pinpointing the development of generative AI specifically, Chadha said technology had reached a powerful “tipping point” in its potential for wide adoption and radical change in both large and small organisations. Taking ChatGPT as an example, he said:

“AI was in the realm of governments, enterprises and large organisations. Today, AI is in the hands of each one of us. That’s what makes it strong”.

In the context of international trade, AI shows a breadth of application across “the entire supply chain cycle”, with the potential to be embedded across “planning, procurement, production, deployment and monitoring” of goods.

One example Chadha offered was the use of “digital twins” to trial the sale of products overseas. AI could facilitate analysis of different international markets’ regulatory and customs requirements, providing insight into the profitability of exporting to new markets without the need for costly consultancy fees.

Smaller actors

Chairing the discussion, Susan Schwab, chair of the board at the US’s oldest trade association, the National Foreign Trade Council, noted that smaller economic actors could also stand to benefit from AI amid increasing regulatory requirements.

She cited farming and manufacturing as examples of industries struggling to keep pace with new regulations mandating the submission of data on human rights, national security, deforestation and carbon emissions within their supply chains.

“You can’t be a manufacturer today – or tomorrow – and certainly not a successful manufacturer, without having a cross-border digital component associated with your manufacturing”.

She added that, because this new regulation creates such a “data-rich environment”, AI could offer considerable support to farmers and manufacturers.

A ‘widening gap’

However, not all panellists were positive about AI’s potential to benefit smaller economic actors.

Asked by Schwab whether AI will be a “a great leveller for small business and small economies” in relation to supply chain efficiency, Torbjörn Fredriksson, head of e-commerce and digital economy branch at UN Trade and Development (UNCTAD), was less optimistic.

“So far, especially with data-driven digitalisation, [AI] has widened the divide.

“Just having data doesn’t do the trick, you need to do something with the data, so it’s very much linked to skills”.

While Fredriksson said that he and many other trade and development experts hope that digitalisation will lead poorer countries to “leap-frog” ahead and “catch up” to their more developed counterparts, he warned there was a risk that it could instead contribute to “enormous divides when it comes to digitalisation”.

Citing research by the International Labour Organization, he said that high-income nations “disproportionately benefit” from AI, while medium- and low- income nations “lag behind” in their use and application of the tech.

He added that within AI research, the academic and commercial organisations producing the top 25 research papers have always been concentrated in the world’s most advanced economies: the US, China, the UK and Switzerland.

To bridge this widening gap, he said it will take “concerted international cooperation” to support smaller, developing countries, adding that there’s not only a risk of increasing the divide but “squandering the potential” of AI to close the gap and create the “widespread social and economic progress” it has the potential for.

Pharma power

Beyond the difference AI could make to businesses, Pfizer’s head of global innovation, trade policy and international government relations, Elissa Alben, highlighted its benefit to consumers.

She explained AI’s role in enhancing pharmaceutical supply chains, many of which led to the unprecedented speed of Covid-19 vaccine development and distribution.

She listed the ability of geographically disparate manufacturing plants to communicate effectively and the greater capacity of the mRNA algorithm – responsible for finding the most efficient way to develop vaccines – as crucial technological developments.

Alben said this helped make it possible to produce 20,000 more vaccines per batch during the Covid-19 pandemic, enabling the total production of 3bn vaccines in 2021 and 4bn in 2022.

“That scale, prior to Covid-19, was unimaginable”.

In terms of vaccine distribution, AI helps to mitigate the disruption caused by natural disasters, facilitating alerts ahead of Hurricane Ian’s landfall in 2022, which enabled Pfizer to intervene and reroute shipments of medicines to trial sites.

Alben also noted the value of AI to another crucial aspect of pharmaceutical distribution: cold chain. Pfizer now uses GPS thermal sensors to track “ultra-cold shipments” of vaccines, ensuring they arrive at remote destination at the correct temperature and therefore usable.

Environmental toll

Fredriksson also noted the risks posed by the rapid development and adoption of AI, namely its high environmental impact.

“In the last few years, AI has added a lot to the energy needs of data centres. This is especially linked to generative AI, which is very compute-intensive”.

These energy needs are made up of large supplies of electricity to power data centres and high volumes of water to cool them.

He shared International Energy Agency research finding that the electricity requirements of global data centres will more than double between 2022 and 2026 – from 460 terawatt hours to over 1000, a major threat to international net zero emissions if renewable energy production can’t keep pace.