International collaboration: the key to developing beneficial AI

We are currently living through an era of major upheavals, which raise equally major societal challenges: pandemics, climatic threats, natural disasters, wars, technological disruptions… To face these global challenges, we have multiplied international collaboration platforms, building places of exchange conducive to the search for common solutions. The development of a technology as transformative and cross-disciplinary as artificial intelligence (AI) is no exception. Many major institutions with international reach have set up AI-related initiatives, such as the OECD’s AI Policy Observatory, the World Economic Forum’s Centre for the Fourth Technological Revolution, the Forum for Cooperation in AI, the International Organization for Standardization (ISO) and the European Commission’s High Level Expert Group on AI, to name but a few. However, a large proportion of the work carried out by these organizations focuses primarily on two aspects: the adoption of artificial intelligence on the one hand, and the creation of standards and public policies surrounding its development, deployment and use on the other.

There’s a little-occupied space in international discussions on AI. A space that the International Center of Expertise in Montreal on Artificial Intelligence (CEIMIA) and the Global Partnership for Artificial Intelligence (GPAI) are trying to fill. The potential of AI for social benefits is currently under-exploited and very little explored. At present, there are little to no business cases for applied AI projects with such vocations. They therefore attract little interest from the private sector, and are not necessarily covered by standardization or public policy projects.

We’ve known for several years that AI has the potential to significantly affect not only our economies, but also our society, our environment, our political institutions and even the way we interact with each other as individuals. Indeed, there have been a number of somewhat alarmist releases recently on the theoretical risks of developing a malicious super AI. But when we leave the theoretical or economic spheres, we rarely talk about the potential benefits this technology could bring to society. And yet, one of the things AI excels at is finding solutions to concrete, complex problems. Major issues such as the fight against climate change, health, the resilience of our societies and respect for human rights are challenges we all collectively face. Stakeholders are slow to address these issues, and since they are unlikely to be tackled by the private sector, it is a shared responsibility to try to find solutions. That’s why we’ve seen increasing calls for international collaboration in this area. The G7 Hiroshima call is a good example of this.

Of course, if we are to bring about change, particularly on global issues such as these, we need to move beyond the discussion stage to the search for solutions. To do this, we need to put in place a structure that encourages dialogue and communication, so that we can identify the real and present challenges. Without such a structure, any project is doomed to be a shot in the dark. Sometimes hitting the mark, but often resulting in a lot of effort deployed in vain. Whether in a commercial or social context, the problem must first be identified before the best solution can be found. This is even truer when it comes to AI, a technology whose effectiveness is highly dependent on the basic premises it is provided with.

As mentioned in the OECD Principles on Artificial Intelligence, international cooperation is needed to ensure the development of trustworthy AI that will benefit society as a whole.

The benefits of greater collaboration between the various stakeholders are numerous and well known:

• Access to diverse perspectives: International collaboration brings together experts from different backgrounds, cultures and industries. This diversity of viewpoints can lead to a more holistic approach to problem solving, resulting in better, more innovative solutions.
• Better knowledge sharing: Collaboration enables knowledge and expertise to be shared, which can lead to faster development of concrete, applied solutions. It can also help to avoid duplication of effort and focus more on the most pressing problems.
• Sharing data and resources: more active collaboration can help overcome the problem of limited access to the databases needed to develop AI-based solutions to tackle global issues. Pooling data from multiple sources will enable larger and more diverse databases to be built up, which can improve the accuracy, robustness and relevance of models.
• Accelerating progress: Collaboration helps accelerate the development of applied solutions by enabling people to work together on complex problems. This can lead to faster and more effective development of high-impact solutions.
• Improving ethical standards and interoperability: International collaboration helps to promote ethical standards and guidelines for the development and deployment of AI. This translates into tools and models developed in a responsible, transparent and socially beneficial way.

As calls for regulation grow louder, so does the need to increase the level of collaboration between organizations addressing the impacts of AI. Technology is developing and deploying faster than any individual government or organization is able to keep up with. The longer we delay working together, the more likely it is that the adverse impacts will outweigh the potential benefits to society.

The GPAI was set up in this spirit of collaboration between states, to concentrate international expertise in a context that favors the emergence of high-impact projects. But we need to go further in our collaborative efforts. The entire international community must unite its efforts and promote constructive dialogue, because no one can do it alone, and the stakes are too high to allow ourselves to be held back by fence-sitting. We have to put aside our differences and our “territorial squabbles”, because technology will not slow down its development, and its impact transcends borders. It’s time to break down organizational barriers to collaboration, and work together in good faith to find common solutions for the good of all.

Picture of Shane Rounce on Unsplash