United Nations in the Era of AI
“Through lines of code, I was able to create an algorithm that, when implemented in society, could actually change people’s lives.” -Stephanie, AI4ALL Graduate
AI is a transformative public good
AI will have an economic impact comparable to the steam engine and could boost global GDP by 1.2% per year by 2030, delivering an additional $13 trillion in economic value. AI is rapidly becoming more adept at performing complex tasks thanks to dramatic increases in data processing speed, storage capability and transfer rates. IBM CEO Ginny Rometty has recently said that AI will impact “100% of jobs, professions and industries.” But beyond these impending economic shifts, the real promise of AI lies in its potential to tackle global challenges like hunger, poverty and climate change.
For example, the Food and Agriculture Organization of the United Nations reports that over 821 million people globally are undernourished. Kimetrica, a social enterprise with bases in the US, Ethiopia and Kenya has developed a machine learning tool called Methods for Extremely Rapid Observation of Nutritional status (MERON) which can instantly detect malnutrition based on a single photo. Especially in remote or conflict-affected areas, technology like MERON could dramatically improve the way that aid is delivered, and could help the UN anticipate emergencies before they break out.
AI is transforming the way we address nearly every global challenge. AI programs are providing real-time monitoring of crop risks to farmers, allowing them to produce more food with lower environmental impact. Using the same tools developed for voice detection, AI programs are now helping seismologists detect earthquakes they might otherwise miss. Across a wide range of medical procedures, AI-driven robots are assisting surgeons, providing more precision and better outcomes.
These are just a few of the potential areas where “moonshots” in AI could lessen the cost of important interventions, prevent disasters, and improve the health and wellbeing of millions of people and our planet.
AI carries huge risks as well
As Christian Lous Lange warned nearly 100 years ago, “Technology is a useful servant but a dangerous master.” As we delegate more tasks and decisions to AI systems, how will humans minimise harm and maximize benefit? Already capable of mirroring human-like abilities, AI is showing the tendency to absorb our own deeply ingrained biases, displaying at times racism, sexism and other forms of discrimination. When AI programs wrongly predict that black prisoners are more likely to re-offend than white ones, or show gender bias in hiring, are they beginning to amplify our own worst natures?
The need to combat bias and other forms of public harm potentially caused by AI is clear. This will require a serious and proactive collaboration across public and private sectors when it comes to data protection and the design of AI programming. But our most urgent imperative—and the step that will address these risks most sustainably—is to prioritize inclusion and diversity from the beginning of the “AI revolution.” Inclusion cannot be an afterthought.
Why does inclusion matter for AI?
Broadening access and inclusion for AI not only mitigates the risks of inequality and disenfranchisement, but it also ensures that AI reaches its fullest potential. Through the organization I lead, AI4ALL, I have already seen the enormous impact that AI inclusion and diversity can yield. In a 2017 program at Stanford, students used AI programming to turn tweets into more efficient disaster relief following Hurricane Sandy. One alumna of an AI4ALL camp has started using machine learning to help radiologists more efficiently detect multiple sclerosis through MRI scans. Another alumna is using AI at an MIT lab to improve decision-making in the US criminal justice system.
AI has given unprecedented access to the top of the job market to this under-represented talent. More than 90 per cent of AI4ALL alumni are confident in pursuing a career in AI, and 97 per cent of them feel part of the AI community. As they branch out, presenting at academic conferences and solving new problems like early detection of wildfires, personalised education and disease diagnosis, they will help broaden the AI field further, benefiting us all.
How do we ensure no one is left out of the opportunity to learn and get involved with AI? One way is to increase AI literacy and access points into the field. For example, our organization recently announced the AI4ALL Open Learning platform: a free, culturally responsive and globally available introduction to AI. Other initiatives are looking to use open-source platforms to give access to anyone wishing to access the AI field.
The fourth industrial revolution can bring prosperity, health and stability to the globe. But it can also accelerate our worst natures, driving marginalization, inequality and destructive growth. By granting access to the revolutionary potential of AI to all of us, we can harness the best of humanity for all of humanity.
This article has been prepared by Tess Posner as a contributor to AI & Global Governance. The opinions expressed in this article are those of the author and do not necessarily reflect those of the Centre for Policy Research, United Nations University, or its partners.