What is AI and why it matters
Hello world! The future is here: how can development practice respond?
4 de Abril de 2023
Since the public release of ChatGPT, an artificial intelligence bot that interacts with users via chat, the topic has gone viral across the world. The AI frenzy has generated many opinions about its potential and limits, its risks, and opportunities, and certainly about the most fundamental questions AI technology poses for society at large.
For decades, technology prophets and futurists announced the arrival of a future in which, on the one hand, humanity may reach its civilizatory peak with unprecedented progress, or collapse under the regime of tyranny. While ChatGPT could be the first-time ordinary people will interact with AI,advances in that field have been underway or decades.
However, the launch of ChatGPT is different: its impact on pop culture, business, and the tech industry appears as if the foretold future has finally arrived, already changing how organizations and individuals engage with digital technologies. So, how can we in international development respond to the widespread use of AI and what it implies for the future?.
In this post, I reflect on AI becoming more mainstream and consider ways in which the international development community might be able to harness its potential for advancing an agenda where everyone is included in its benefits.
What is AI and why it matters?
A concept that encapsulates the general understanding of AI describes it as “machines that respond to stimulation consistent with traditional responses from humans, given the human capacity for contemplation, judgment, and intention”.
An important caveat is that AI is not a single technology, but a range of technologies advancing at different paces and levels, from content generation (e.g., text, video generators) to robotics (e.g., drones) to smart language, and speech processors, each with different levels of human intervention and purposes.
While no single entity dominates AI, a handful of tech giants are leading the race, which raises questions about data privacy and ethical considerations about the reach and scope of AI into people’s private lives.
New technologies, old issues: AI and development’s points of divergence
The most recent advances in AI may give us reasons to be cautiously optimistic about its potential. Even so, for us in the development sector, it is important to be reminded of old issues the world still confronts:
• The future of work and people’s livelihoods: One of the most somber predictions about AI is automation and its impact on labor market dynamics. The fear is that millions of jobs could be replaced on a never-seen-before scale. The world is still grappling with issues such as informality and precarious livelihoods, low wages, job insecurity, gender pay gap, etc. Will AI deepen the problems or help fix them?
• Digital exclusion: ITU statistics show that only 66% of the world’s population have access to the internet. The rest, about 2.7 billion, "are still offline". A World Bank report also says that "some 850 million people globally do not have an official ID". How could AI’s benefits spill over to those who are not yet part of the digital economy? How can the global south move from consumer to producer of AI technology?
• Global inequality on the rise: Although the world has become richer and more prosperous, large chunks of its population struggle with bread-and-butter issues. The widening of socioeconomic gaps threatens social cohesion and political stability. The UN’s World Social Report reminds us, "Countries where inequality has grown are home to more than two thirds (71 percent) of the world population".
• Governance, identity politics and polarization: research suggests that AI can play a key role in fueling polarization and identity politics across the globe, with algorithms designed to spread misinformation because that generates more engagement. Ideological radicalization is on the rise, making it more difficult for political compromise on core development issues.
The above may suffice to underscore the need for a global compromise on a human-centered development of AI.
Development and AI: How UNDP see it
Through the Digital Strategy from UNDP, lays out how we support countries to build inclusive, ethical, sustainable digital societies. This implies taking a holistic approach that integrates people (digital skills, digital well-being), physical infrastructure, governance in all its structures, regulations that must accompany these changes, and the private sector. This strategy also recognizes the potential of artificial intelligence and emphasizes that ethical and responsible use of technology for the benefit of society should be the norm.
Furthermore, the UNDP Global Gender Equality Strategy 2022-2025 establishes that increased access for women to digital technologies, digital finance, e-commerce, digital value chains, and their utilization are central elements to eliminate gender discrimination, occupational segregation in labor markets, reduce informal and precarious work, and increase women's employment rates. For instance, between 2017 and 2018, 35% of women, compared to 24% of men, reported not knowing how to use a smartphone in the region, and 40% of women versus 33% of men reported not knowing how to use the internet.
While AI and human development may have separate agendas, there are points of convergence where there is great potential for social progress.. An IDIA report says that “AI demonstrates significant potential to solve some of the most pressing issues facing society”. Thus, a task for development organizations is to learn how to navigate this new terrain. Here is why:
Here is why:
1) AI may provide tools designed to pinpoint more accurately and efficiently the development issues we try to address.
2) Innovation has been permeating the international development ecosystem in recent years, normalizing the space for a flexible, experiment-oriented approach to project design and implementation where all stakeholders can engage more meaningfully in testing potential solutions. AI’s innovations can be catalytic to project success.
3) Expand interventions’ scope: whether it’s agriculture, healthcare, or education, by incorporating emerging, groundbreaking applications, AI can strengthen initiatives’ impact, reducing guesswork, manual labor and increasing productivity and effectiveness. More importantly, we can make AI accessible to populations who otherwise may not have access to it.
Development practice is still relevant
In the face of these developments, if we are to keep up with social change, we can no longer do development as usual. As practitioners, we can function as equality brokers, sensing where innovations arise and striving to make them useful for tackling development challenges.
At UNDP/DR, we have been working with AI and digitalization initiative and partners that show promising results. As an example, in the framework of the regional project Infosegura, UNDP in partnership with USAID are developing SARA, an AI/machine learning chatbot assistant for women facing domestic violence, a service being implemented along the Ministries of Women and Police. That initiative aligns with UNDP’s global DigitALL campaign, a call for prioritizing the use of innovation and digital technologies in the pursuit of gender equality.
In the context of the work carried out on the project “Rescate Ozama” on the Ozama River, The Ocean Clean Up, a strategic ally for UNDP in this initiative, is developing cameras powered by Artificial Intelligence that are located on the bridges, which automatically detect how much plastic is floating through/down the river.
UNDP/DR’s lab and its allies, partnered with and co-financed (via Digital X) eHospital, a telemedicine project which used Pamela, a chatbox doctor capable of assessing and providing directives to patients as for what to do during the COVID pandemic. eHospital later was scaled to provide teleconsultations in a remote DR community. Similarly, UNDP’s Digital X featured the AIME Console, an AI platform that uses data to predict epidemiologic outbreaks. It is quite evident that AI can be a powerful enabler of solutions that benefit society’s most vulnerable.
AI learns more from humans every day, and therefore, it can also learn to replicate our biases. That's why its development must be based on ethical principles, centered on people on equal terms, transparency and accountability, as well as security and privacy, with the aim of maximizing its potential for sustainable development without leaving anyone behind.
AI algorithms are not by themselves good or bad; they have no feelings, no empathy for the poor and the vulnerable. But we do. Development is still relevant because our work is a humane endeavor, where people care for other people to make a better society. And because humans create technology, it is our responsibility to ensure whatever technology is produced improves the human condition.