AI at the service of the common good

LONDON – The tech world generated plenty of front-page news in 2022. In October, Elon Musk bought Twitter—a leading public communication platform for journalists, academics, businesses, and governments—and proceeded to lay off most of the moderation staff. of content, indicating in doing so that the company was going to use artificial intelligence (AI) instead.

Then, in November, a group of Meta employees revealed that they had designed an AI program capable of beating most humans in the strategy game Diplomacy. In Shenzhen, China, the government is using “digital twin” technology with thousands of 5G-connected mobile devices to monitor and manage the flows of people and cars and energy consumption in real time. And with the latest version of the ChatGPT predictive language model, many have declared the end of the student thesis.

In short, it was a year in which serious concerns regarding the design and use of technologies deepened to give rise to even more urgent doubts. Who is in control? Who should have it? Policies and public institutions are needed to ensure that innovations imply improvements for the world; but today, many technologies are developed in a vacuum. We need governance structures with a sense of mission that are inclusive and centered on a true common good. Governments equipped with the necessary capabilities can guide this technological revolution so that it serves the public interest.

Take the case of AI, which the Dictionary of the Royal Spanish Academy defines as follows: “Scientific discipline that deals with creating computer programs that execute operations comparable to those performed by the human mind, such as learning or logical reasoning.” AI can bring us numerous benefits; For example, improve food production and management by increasing the efficiency of agriculture and the safety of food products, or help us strengthen resilience against natural disasters, design energy-efficient buildings, improve energy storage and optimize the deployment of renewable energy sources. And it can increase the accuracy of medical diagnoses (when combined with professional opinion).

These apps promise a variety of benefits. But without effective rules, there is a risk that AI will create new inequalities and amplify existing ones. You don’t have to look far to find examples of AI-based systems that reproduce unfair social biases. In a recent experiment, robots driven by a machine learning algorithm displayed overtly racist and sexist behavior. Without proper supervision, algorithms supposedly designed to help the public sector manage social benefits may end up discriminating against families in real need.

Another equally worrying example is that in some countries, public authorities are already using AI-based facial recognition technologies to control political dissidents and subject citizens to mass surveillance regimes.

Another important problem is the concentration of markets. A few geographically concentrated powerful players dominate AI development (and control of the underlying data). Between 2013 and 2021, 80% of global private investment in AI was made in China and the United States. There is already a huge power imbalance between the private owners of these technologies and everyone else.

But AI also receives a huge amount of public funding, and that funding should serve the common good, not the few. We need a digital architecture that more equitably distributes the benefits of collective value creation. You can no longer have a non-interventionist position based on self-regulation. Giving free rein to market fundamentalism is condemning the State and taxpayers to repair the damage later (as we have seen in the context of the 2008 financial crisis and the covid-19 pandemic), not without great financial costs and scars. lasting social In the case of AI it is even worse, because we don’t even know if an ex post intervention will be enough. As The Economist recently noted, it’s common for even AI developers to be surprised by the power of their creations.

Happily, we already know how to prevent laissez faire from causing a crisis again. We need a mission to create AI that is “ethical by design”, based on strong regulations and the work of governments equipped with the necessary capacities to guide this technological revolution in the direction of the common good and not only for the benefit of shareholders. . Once these pillars are in place, the private sector will be able and willing to join the broader effort to create safer and fairer technologies.

Effective public oversight is needed for digitization and AI to create opportunities for public value creation. This principle is an essential component of the Recommendation on the Ethics of AI, a normative framework proposed by UNESCO and approved by 193 member states in November 2021. In addition, key actors have begun to take responsibility for reframing the debate; In the United States, the Joe Biden administration has proposed an AI Bill of Rights, and the European Union is developing a holistic framework for AI governance.

But the public use of AI must also be based on a solid ethical foundation. Faced with the advance of this technology as an auxiliary tool for decision-making, it is important to avoid using AI systems that are contrary to democracy or human rights.

The lack of investment in public sector innovation and governance capacities also needs to be addressed, which (as COVID-19 made clear) need to be much more dynamic. If, for example, clear terms and conditions are not set for public-private partnerships, there is a risk that companies will take over the agenda.

But the problem is that the subcontracting (outsourcing) of state functions has become a major obstacle to capacity building in the public sector. To maintain control over important products and ensure ethical standards are respected, governments must be able to develop AI without relying on the private sector to provide sensitive systems. They also need to be able to sustain information sharing and interoperability of protocols and metrics between various departments and ministries. All of this will demand public investment in the capacities of the State, according to an approach based on the idea of ​​mission.

Given the concentration of knowledge and experience in the private sector, the creation of synergies between the public and private sectors is both inevitable and desirable.

The sense of mission has to do with selecting actors willing to collaborate, through joint investment with partners that recognize the potential of state-led missions. The key is to give the State the ability to manage the deployment and use of AI systems, instead of always running behind events.

One way to ensure a correct distribution of the risks and benefits of public investment is for governments to set conditions for the provision of public funds. They can also, and should, demand more openness and transparency from big tech companies.

The future of our societies is at stake. We must not only correct the problems and control the risks of AI, but also influence the general direction of digital transformation and technological innovation. What better time than the beginning of a new year to start laying the foundations for a process of unlimited innovations at the service of all humanity?

*Gabriela Ramos is Assistant Director General for Social and Human Sciences at UNESCO and Mariana Mazzucato is Founder and Director of the Institute for Innovation and Public Purpose at University College London, Chairs the Council on the Economics of Health for All of World Health Organization.