Decoding AI: Potential, Pitfalls, and Promises in International Development
In the dynamic tapestry of modern technological innovation, Artificial Intelligence (AI) stands out as a weaving thread, connecting the dots between data, decisions, and development. As AI transitions from the realm of sci-fi to real-world applications, delve into its transformative role in global development. Explore how, from healthcare innovations to climate resilience strategies, AI is redefining the possibilities for a brighter, more sustainable future.
There is a moment in the life of a technological innovation when it moves from being something peripheral to becoming embedded in our lives. Like the Internet after the launch of web browsers, or social media after Facebook. Much like the pre-history of the world wide web, Artificial Intelligence has been around in various avatars for decades. But in the past few months the term has acquired a new hold over the public imagination. Suddenly, AI systems seem increasingly ubiquitous, from being used in healthcare and financial settings, to planning out our everyday chores and routines.
With this degree of presence, it can be easy to forget what exactly constitutes Artificial Intelligence. AI is defined as “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.” AI systems process and learn from large data sets, create algorithms and models, and make predictions or decisions based on their understanding of the information they process. Crucially, AI can sift through massive amounts of data, far exceeding human capacity. It can identify subtle patterns and correlations that might not be evident to human analysts. And it can do all this quickly! In this way, AI can solve complex problems, minimise waste and inefficiency, and offer insights for future planning.
The transformative nature of AI regarding fake news and disinformation in particular is massive, and has sparked both concern and hope across different quarters.”
Specifically for the development sector, AI seems to represent a continuation of the long-held belief in the transformative potential of technology. For decades, planners and policy makers have looked to technological innovations to transform the lives of citizens and create a better future for large parts of the planet. Consider, for instance, how low cost mobile phones have helped vulnerable communities in Kenya access mobile payment systems through the M-Pesa platform. During the pandemic, ASI’s partner organisations in Somalia managed to establish information campaigns that dealt with questions and rumours regarding COVID-19 through 1-to-1 messaging. They were able to design their information campaign keeping in mind the specific pieces of misinformation or gaps in knowledge that they could identify through their SMS communications and social media platforms. Similarly in Pakistan’s Khyber Pakhtunkhwa region, the Sustainable Energy and Economic Development (SEED) Programme helps provide business and digital skills training for home-based women entrepreneurs. The crux of this programme is to take on the gender gap in economic activity in the country. Pakistan ranked 148 out of 149 countries on the World Economic Forum’s Global Gender Gap Index in 2018, with just 22 percent female participation in the labour force. The country also has one of the lowest rates of women entrepreneurs in the world. The SEED programme uses the internet and technology to support home-based businesses in Pakistan, allowing women to work and earn from home, paving the way for long term socio-economic change.
AI is the latest turn in this long running, global story of connection between human innovation and the desire for a better life. Can it deliver on its promise, more so than other breakthroughs before it?
Even a cursory look at the sector shows the diverse ways in which AI-based innovations are already being implemented. For instance, UNICEF is using facial recognition technology to detect malnutrition in children, through an algorithm that analyses a child’s photograph to estimate body mass index. In the Philippines, researchers have used machine learning tools to map potential outbreaks of dengue fever with increasing accuracy, using weather and land use patterns connected to transmission of the illness. And cities like Dublin and Singapore have created their “Digital Twins” to replicate their particular environments, and use AI to predict and solve future problems.
In Pakistan, ASI has used a similar approach in assisting the KP government in building climate-resilient infrastructure. Rather than using the more traditional methodology of relying on historical data, we developed a set of complex computer-driven simulation models to create climate forecasting in certain flood-prone regions. The team employed advanced climate modelling techniques, augmented with AI-based analysis, along with primary field data to forecast future climate risks in target districts since historical data alone can no longer be held as reliable in the face of unexpected changes caused by climate change. The results of this forecasting model over 100 and 200 year time periods show the impact of climate change to be much higher than anticipated. The greater risk of flooding, for instance, needs to be take into consideration while planning major infrastructure systems. Such information is essential to build resilience to the potential catastrophes caused by climate change in the future, and to offset the costs of climate inaction now.
So what can be done to address this paradox?
The Oxford Insight’s Government AI Index 2022 states that “(AI) is increasingly part of governments’ plans to reform public services… However, there is a lack of understanding about precisely what foundations are needed for a government to be in the position to integrate AI into services and, beyond that, what it takes for AI to then be used in government effectively and responsibly.” Legislation to ensure responsible use and preparing frameworks for accountability are thus key inputs required from governments, especially in low income countries. Ethical guidelines are essential. Governments must work with experts to develop policies ensuring fairness, transparency, and accountability in AI systems.
Side by side, digital literacy initiatives aimed at creating an inclusive approach can foster more equitable use of technology for development. As we have seen in Somalia and in Pakistan, investment in technical training for women and marginalised communities helps diversify the power for decision making.
The advent of generative AI, said António Guterres, Secretary-General of the United Nations, “could be a defining moment for disinformation and hate speech”. The UN has also noted that it has the “potential to turbocharge global development and realise human rights” but can also “amplify bias, reinforce discrimination and enable new levels of authoritarian surveillance.”
The significance of AI in countering disinformation and potential radicalisation is becoming increasingly pivotal in the development sector. While maintaining user privacy and adhering to ethical guidelines, AI can scour vast digital landscapes, identifying and flagging content that is intended to misinform or radicalise. For instance, machine-learning algorithms can analyse patterns, contexts, and sources of information to differentiate between genuine and potentially harmful content. Particularly in regions grappling with disinformation and radicalising materials, these AI systems can work towards mitigating the spread of such content by identifying it in real-time, thereby enabling timely interventions. This technological stride not only provides a mechanism to combat the pervasive issue of fake news and extremist content but also aligns with global efforts to stabilise regions affected by disinformation-driven conflict, ensuring a safer digital space for vulnerable populations.
It is not just an urgent necessity but a pathway that may already exist, and can be built on using innovations, multi-sectoral collaborations and ethical guidelines. It is a road based on effective partnerships and collaborations between governments, international agencies and experts, as well as community organisations. At the heart of each such measure, is the idea of technology creating a better life for the most vulnerable.
As AI embeds itself deeper in our world, fostering inclusive AI development is an essential part of our imagined future.”
ASI’s In-depth Study in KP
In a collaborative effort between ASI and the Government of Khyber Pakhtunkhwa (GoKP) under the SEED initiative, an innovative approach was taken to address climate change impacts in the KP region of Pakistan.
The Challenge: Traditional methods, which often rely on historical data, aren’t as effective in today’s rapidly changing climate landscape. The unpredictability of climate change necessitates more advanced forecasting techniques.
Our Approach: We turned to cutting-edge computer-driven General Circulation Models (GCMs) in tandem with complex statistical algorithms for climate forecasting, especially focusing on flood-prone areas.
Key Insights from the Study:
- Climate Variability: Climate projections showed a range of possible futures, with temperature potentially peaking by up to 5.7°C by the end of this century, shedding light on the urgent need to prepare for various climate eventualities.
- Incorporating Climate Data into Infrastructure Planning: Our analysis underscored the importance of embedding versatile climate data in the strategic planning and design of vital infrastructure, safeguarding against elevated flooding risks.
- Future-Proofing KP: By understanding and integrating climate change profiling in activities like flood plain mapping and adhering to local regulations like the KP River Protection Ordinance 2002, we can better prepare for future challenges.
- Climate Risk Checklist: As part of our recommendations, we introduced a checklist for the public project appraisal process, helping authorities ensure they’re considering all potential risks linked to climate change.
- Policy Options for Robust Climate Strategies: Emphasising the imperative of actionable policies, the study highlighted a need to explore policy options that ensure climate-resilient development. This includes investing in precise climate modelling, enhancing data collection, and cultivating research and capacity-building initiatives. Additionally, policies should facilitate international collaboration and comprehensive long-term planning, intertwining nature-based solutions to ensure sustainability amidst climatic challenges. These strategies collectively forge a path towards a future adept at navigating through the multifaceted impacts of climate change.
Why It Matters: While the study leveraged advanced climate modelling techniques andmachine learning (a subset of AI), it wasn’t even fully realised using AI – and yet, its implications for preparedness are profound. Imagine the potential accuracy and actionable insights we could garner with full AI integration, optimising our strategies further. These insights aren’t just academic exercises; they’re crucial in helping us prepare for and mitigate the potential catastrophic effects of climate change. Proactive measures today will save costs and lives in the future, ensuring the KP region (and other countries and regions vulnerable to climate effects) is resilient and prepared.
Across the world, states are working to create regulatory frameworks for AI. As the technology itself undergoes quick changes in short periods of time, many of these laws and regulations are also in flux. The first such regulatory framework was proposed by the European Commission in April 2021. The crux of the rules is to ensure that AI systems should be overseen by people, rather than by automation, to prevent harmful outcomes. The AI Act, once approved, will be the world’s first rules on AI and seeks to set global standards in the sphere, similar to the GDPR’s (General Data Protection Regulation) impact on individual privacy.
According to the Economist Intelligence, China’s ambitions to be the be the global AI leader are balanced by the need of its ruling party to exercise control over AI systems. The April 2023 draft regulation on generative AI thus will “focus on responsibility, privacy, accuracy and misinformation, and this makes them the closest to the EU rules anywhere in the world.” The rules will however have less of a global impact as they will be applied in a very different political context.
The United States faces a challenge from its political system while creating regulatory frameworks, while also favouring innovation and self regulation over control.
Disclaimer: This article has been written by a human.
We did, however, ask an AI what IT thinks how it can best support governments, especially those in complex or developing contexts and the below response in the below is what it said.
AI’s Insight on Empowering Governments:
When pondering the multifaceted ways AI can strengthen governmental strategies, especially in fragile or emerging economies, the potential is vast:
- Data Processing: Harnessing AI’s capability to decipher vast and complex data sets can guide policy-making and strategic governance decisions.
- Predictive Analysis: From anticipating disease outbreaks to forecasting economic trends, AI’s predictive algorithms allow governments to stay several steps ahead.
- Operational Efficiency: AI can optimise public service delivery, from traffic management to utility distribution, ensuring resources are utilised effectively.
- Strategic Investments: AI can analyse market trends, demographic shifts, and societal needs to suggest areas that would benefit most from investment.
- Revolutionising Education: Beyond generic content, AI can curate bespoke educational material that caters to regional and cultural nuances.
- Security and Surveillance: AI-powered systems can aid in crime prediction, border control, and overall national security enhancements.
- Disaster Management: Predictive modeling can anticipate natural disasters, and post-disaster, AI can aid in efficient resource allocation and recovery efforts.
- Public Health: AI can analyse patterns in public health data, helping to design health campaigns or detect early signs of epidemics.
- Agriculture & Food Security: Through AI, governments can predict crop yields, detect potential infestations, and optimise irrigation, ensuring food security.
- Sustainable Urban Planning: AI tools can be used to design cities that are more sustainable, energy-efficient, and responsive to the needs of their inhabitants.
- Public Engagement: AI chatbots and platforms can facilitate better communication between governments and citizens, addressing queries and gathering feedback.
- Economic Forecasting: Governments can use AI to predict economic booms and downturns, adjusting policies accordingly.