Shenzhen, October 2018: Like most tech hubs in the city, the Shenzhen Open Innovation Lab feels busy, modern, thriving. In the lab, you meet mash-ups of self-made engineers, inventors, and artists working on projects ranging from drone swarming for diagnosing diseases on crops to precision algorithms for companion robots. Shenzhen made a simple, winning bet: hacking the future.

Shenzhen’s community of do-it-yourself entrepreneurs is increasingly sharing ideas and designs with engineers at Kumasi Hive, a collaborative incubator for entrepreneurs in Ghana, thousands of miles across the ocean. With agile tech infrastructure and legacy regulation, hubs in Africa are like Shenzhen thirty years ago. These democratised innovation ecosystems are ready to learn, adapt, and harness artificial intelligence (AI) for their own needs.

Some words — like those of William Gibson — stay with us through our journey. “The street finds its own use of things.” Inspired by that mantra, Shenzhen experiments with “AI from the streets”, giving diverse communities around the globe an opportunity to turn their data, ideas and designs into AI innovation. Such a diversity of knowing and experiencing the world is the best remedy to help foster diligent technical design, anticipate failures, and minimise the risks of unintended harm.

Open AI and local dynamism

Mainstream discussions of AI and robotics assume that a small number of global tech firms will control the technologies that will affect the lives of massive populations. Recently, bitter criticisms have emerged of large private tech platforms that favour profits over public interest, undermining public trust. As a result, strategic discussions in AI governance circles have increasingly focused on defining ways to restrict and regulate the large corporations that lead the AI industry.

There is an alternative path for developing and deploying AI, what proponents of open innovation call “AI from the grassroots”. In China, democratised ecosystems, which build on open source approaches to new technologies, have developed surprisingly active local dynamism with incentives to better connect tech with real social issues. Increasingly, AI should be considered as a tool of what Edmund Phelps has coined “mass flourishing”. Nations or regions can truly develop, implement and grow with new technologies only if they maintain a regulatory environment that enables technological transfer to the grassroots level, delivering valuable products and services not only to the wealthy and powerful but also to social and economic peripheries.

Why democratise innovation?

AI naturally lends itself to this kind of local, decentralised approach. Knowledge and means of technological production — the two key ingredients for innovation today — are available to anyone with internet access. Most AI technologies can be obtained via open source licenses. Many core learning algorithms are available on public platforms. In China, for example, the means to build the kind of devices that run AI on smartphones, robots, and self-driving cars are all available within the open innovation and production ecosystem.

Home-grown innovators around the world can now leverage global resources and tailor tech production to suit local needs, developing a knowledge advantage over their global competitors. As local entrepreneurs and users start experimenting with AI in democratised ecosystems, the ability to produce bespoke, responsive products will increasingly favour these grassroots innovators.

Shenzhen as a global hub

The evolution of Shenzhen into a global innovation hub vividly shows the path towards localised technology production. Thirty years ago, Shenzhen was a collection of fishing villages with a population of less than 200,000. Today, following a government initiative to turn it into a special economic zone, it is home to 15 million people and a massive hub for tech exports worldwide. What was previously a small production facility for low-quality industry has rapidly been transformed into a one of the largest and fastest prototyping hubs in the world.

Though a locally driven enterprise, Shenzhen has massive global significance. Apart from Apple and Samsung, nearly every other major smartphone in the world is produced by Shenzhen manufacturers, including Huawei, Vivo, Oppo and hundreds of others. The scale of the electronics industry in Shenzhen has surged from about US$1.59 billion in the 1990s to about US$291 billion today.

With its strong manufacturing ecosystem embedded in the global supply chain, and its open innovation culture emphasizing knowledge sharing, Shenzhen is uniquely positioned to contribute to a future where AI-enabled smart products respond to the needs of diverse populations. The firms in Shenzhen have decades of experience designing, prototyping and assembling products for multinational corporations. Over time, they also learn to excel in researching consumer needs, creating new products with rapid prototyping, and adapting the technology based on local resources, first in China, and now increasingly in other underserved emerging markets, like South-East Asia and Africa.

From Shenzhen to Kumasi

Shenzhen is increasingly becoming a model for innovators in Africa.

Using local data when training algorithms, Tecno — the number one mobile phone company in Africa — has developed facial recognition software tailored to the African market. Tecno’s facial recognition system helps provide an alternative to the Western-biased algorithms that currently dominate the market.

Trotro, a Ghanaian company, experiments with AI-driven global positioning system trackers, produced in Shenzhen, to enable farmers to optimise scheduling. The Ethiopian government recently launched a “Designed in Ethiopia” competition which explicitly plans to “adapt the Shenzhen open innovation model to Ethiopian economic and industrialization processes”.

What policy for democratised AI?

The global innovation ecosystem could be slowly transformed if Shenzhen and other hubs continue to export their democratised approach to AI and new technologies. The future of AI could be defined, invented, and implemented by bottom-up networks of inventors and engineers, rather than large corporate platforms alone.

Several AI and policy experts have voiced the need to build a social license for AI, including new incentive structures to encourage state and private actors to align the development and deployment of AI technologies with the public interest.

In this context, the UN and its Member States, the World Bank, and international foundations, should discuss how to empower and oversee democratised innovation ecosystems, the “grassroots”, in their effort to design and deploy AI for solving local social problems. The United Nations Secretary-General’s High-level Panel on Digital Cooperation could foster inclusive forums of engagement in which to share lessons learned, exchanges of innovation and oversight practices for AI “mass flourishing”.

By empowering “AI from the streets”, our cities could become globally connected, yet locally inventive and inspired by a diversity of knowledge and vision about our shared digital futures.

Echoing Gibson, the future is already here. But we need to distribute its promises more evenly.

•••

David Li and Eleonore Pauwels wrote this article as contributors to AI & Global Governance, an inclusive platform for researchers, policy actors, corporate and thought leaders to explore the global policy challenges raised by artificial intelligence. The opinions expressed in this article are those of the authors and do not necessarily reflect the opinions of the United Nations University.