Pasar al contenido principal
 

Artificial intelligence (AI) is poised to enhance productivity and innovation around the world. The expected benefits promise to be transformative, but the negative repercussions could be magnified in developing countries, where the livelihoods of many people are precarious and social institutions can be fragile.

AI’s influence will be widespread because it can be integrated with other technologies and applied to almost any activity that involves information and communication technologies. In an effort to improve understanding of how to ethically and equitably implement AI in the development context, IDRC has published the white paper Artificial intelligence and human development. It outlines the potential benefits and risks of this new technology and presents a proactive research agenda to address challenges posed by AI that are of particular concern in the developing world.

Learn more about AI and read concrete research recommendations in the full IDRC white paper: Artificial intelligence and human development: Toward a research agenda (PDF, 36.6MB).

Read more on our work in AI in a featured UNESCO blog post.

AI and the potential for development

AI is an area of computer science dedicated to creating software that can be taught to perform complex procedures. What makes AI “intelligent” is that it can learn new behaviours, improve performance as more experience is gained, and make decisions and predictions based on available data. The algorithms at the core of some AI systems are trained using the large datasets that are now available thanks to the “big data” revolution. It is the intelligent capabilities of AI systems that allow for the automation of tasks that until now required human judgement to deliver.

There is enormous potential for how AI can benefit the developing world and what it can contribute towards achieving the UN’s Sustainable Development Goals:  

Healthcare 

AI can play a crucial role in augmenting healthcare capacity by filling gaps in human expertise, increasing productivity, and enhancing disease surveillance.

Agriculture 

AI applications can provide critical insights and solutions to improve the efficiency and quality of agricultural activities. AI is already being used to support water management in the Middle East and drought monitoring in Africa. Farmers in Uganda have access to mobile phone-based tools for the automated identification and tracking of crop infestations.

Economic development 

AI has the potential to drive growth through innovation, increased productivity, and the optimization of business processes. Start-ups in several African countries are building a variety of innovative businesses based on AI systems. Other companies are using AI-enabled mobile phone platforms to provide access to financial services for hundreds of millions of Africans who either do not or cannot access these services through traditional banks.

Education and training 

AI techniques can be used to support the roles of teachers, tutors, and administrators by providing personalized learning opportunities at scale, such as the intelligent tutoring systems currently under development in India.

Greater government efficiency and transparency 

The delivery of government services and information could also be improved using AI systems — ideally to maximize social returns while minimizing financial cost. AI systems can provide automated access in multiple languages and dialects, and high-level decision-making could be enhanced by automating complex assessments that incorporate a range of technical, organizational, and social factors.

Recognizing AI’s risks and challenges

As with any widely adopted technology — especially one as powerful and potentially pervasive as AI — the benefits come with risks that must be managed and mitigated. Many examples have already been documented of these risks playing out in real-world applications.

Exacerbating societal biases 

AI systems can reflect societal biases and assumptions held by their designers or inherent in the datasets on which their core algorithms are trained. Using biased systems to automate decision-making processes could amplify the impact of these biases by systematically producing results that disadvantage particular individuals and groups, especially those who are marginalized. As an example, a computer program used in the U.S. to assess the risk of re-offense by individuals in the criminal justice system was shown to flag black defendants as high risk nearly twice as often as white defendants.

Threatening privacy 

The use of AI algorithms can supercharge the capacity for surveillance, and thereby threaten privacy. For example, by integrating AI-powered facial recognition software, closed-circuit TV systems can track individuals as they move through the urban landscape. Such blanket monitoring is concerning and could be exploited both socially and politically. Unchecked surveillance has the potential to erode privacy, which is key to other fundamental rights such as freedom of expression and association.

Loss of jobs 

With the growing use of machine learning and AI systems in nearly all sectors of the economy, widespread automation will extend beyond manufacturing to impact knowledge-based roles. Many of these positions can be partly or entirely automated, reducing the need for human workers. Estimates of the extent of job losses due to automation vary greatly, but it is expected that the pace of change will be rapid, giving societies and governments limited time to adjust​

Fake news and misinformation 

In a highly connected world reliant on online sources of information, misinformation is a genuine and growing threat to stability and democracy. By capitalizing on the vast amounts of personal data collected via social media, AI applications can facilitate and automate far-reaching propaganda and behavioural manipulation campaigns. The 2016 U.S. presidential election has become a notorious example of the role of targeted misinformation.

As the above examples illustrate, there is potential for AI to increase inequality and generate economic disruption, social unrest, and even political instability in the developing world. While these risks are only beginning to be addressed by even the most advanced countries and economies, the developing world faces particular contextual challenges. The governance and regulation of technology are rudimentary in many countries in the Global South. Communication and digital services infrastructure is lacking in lower-income and rural areas, and there is a general skills shortage for the development and deployment of AI applications. Furthermore, a significant comprehension gap between social scientists, policymakers, NGOs, and those with technical understanding of AI will make addressing these challenges problematic.

Learn more about AI and read concrete research recommendations in the full IDRC white paper: Artificial intelligence and human development: Toward a research agenda (PDF, 36.6MB).

Top image: The Roaming Platypus