Artificial intelligence has found many applications in various industries. Most recently, a field that has been growing both in popularity and traction is applying AI for social good, or AI4SG.
Although AI is not the panacea for the world’s most challenging social and environmental problems, it can offer resources for tackling some pressing issues and deliver positive outcomes.
Many companies have started collaborating with non-profit organizations dedicated to solving some of the global social, economic and environmental challenges or have invested in initiatives that utilise AI to improve social diversity, equity, inclusion and sustainability.
The Tech Giants and AI for good initiatives
If organisations can take a good example of AI for social good from someone, that is the tech giants and their AI4SG initiatives.
The world’s tech leaders have whole business initiatives where they invest resources in building AI solutions for social good or help organisations that are working on positive social initiatives.
For instance, Google AI For Social Good works on addressing some of the biggest social and environmental challenges such as developing systems for forecasting floods, helping people with atypical speech be better understood, predict cardiovascular problems by eye scan, and Google Research India’s AI4SG initiative that supports organizations from India and across Asia focusing on addressing social, humanitarian and environmental challenges with AI.
Microsoft collaborates with an India-based company to eradicate preventable blindness. The company founder K Chandrasekhar established the company with the same goal, but his biggest challenge was the expensive and rare devices for detecting diabetic retinopathy. He reached out to the tech giant as he knew that AI could help create affordable and portable eye-scanning devices. As Microsoft’s AI assists doctors in giving a faster diagnosis, they are able to make a difference on a larger scale, especially in a country in India that has an exponential rise in diabetic retinopathy.
IBM’s Science for Social Good initiative helped stop the spread Zika Virus in collaboration with the Cary Institute of Ecosystem Studies. The team of researchers used two models – multiple imputation and Bayesian multi-label machine learning to assign primate species (likely to harbour Zika and potentially transmit the virus) with a risk score indicating their potential for Zika positivity. The machine learning models identified carriers with 82 per cent accuracy and helped in creating an interactive map that identifies hotspots where people are most at risk of the virus.
Facebook has also launched many initiatives for AI for social good, particularly focusing on promoting best practices of AI for promoting social good in India. One of the most significant projects was a blood donations tool to help connect people interested in donating blood with local hospitals and blood banks, and help people easily locate events, camps and requests for blood donors. The blood donation feature quickly caught on with more than 35 million people who signed up to be blood donors. It grew into a complex technical challenge that only AI can solve. When someone posts on Facebook related to donating blood, AI recognises the content and automatically sends the person or organisation a message notifying them about the feature and inviting them to participate in it.
AI for social good applications for solving global challenges
Apart from the tech giants and their AI for good initiatives, there are also numerous examples where AI is applied to tackle the most pressing social and environmental challenges.
Fight against COVID-19
We’ve seen how IBM’s initiative helped halt the spread of the Zika virus. In the most recent case, AI for social good was implemented to fight the COVID-19 pandemic. As a global pandemic with many variables and unknown traits of the virus, modelling the COVID-19 virus to stop its spread proved extremely difficult. To help the community of epidemiologists, analysts and researchers address the health and economic impacts of the virus, teams from Google with the help of Google Cloud have developed a COVID-19 Open Data repository — a comprehensive, open-source resource of COVID-19 epidemiological data and related variables like economic indicators or population statistics from over 50 countries.
Crisis situation response
AI can also help in many crisis situations such as responses to natural and human-made disasters in search and rescue missions. In 2020, Stanford researchers developed a deep-learning model that maps fuel moisture levels for better fire predictions during the wildfires that raged across the Western U.S. states. Their models leveraged recurrent neural network, an artificial intelligence system that can learn to recognise patterns in vast mountains of data. The researchers trained the model on 3 years of data for 239 sites across the American West. The model output was put into an interactive map that helped fire departments to prioritise response.
Environmental challenges and climate change
In the last several years, we have been witnessing stronger signs that climate change is getting worse. Experts warn that if we don’t take serious actions now, the environmental and economic damages caused by climate change will be irreparable. The zero-carbon transition requires measures that reduce greenhouse gas emissions and build resilience towards weather-related disasters. Luckily, promising AI applications have been developed that help utilise the troves of data generated from different sectors to optimise how we use natural resources.
More specifically, AI can help with impact and resource decoupling, which means decreasing environmental harm, including CO2 emissions, per unit of economic output. Some of the ways AI can help do impact decoupling this are:
- Forecasting the supply and demand of power in the grid, improving the scheduling of renewables, and reducing the life-cycle fossil fuel emissions through predictive maintenance.
- In transportation, AI can enable more accurate traffic predictions and optimisation of freight transportation, to help model demand and shared mobility options.
- In managing the waste and pollutants that affect human and animal health and destroy biodiversity, AI can help make better climate change predictions in order to steward these ecosystems.
AI also plays a role in resourcing decoupling, meaning decoupling of economic output from the volume of resources used from the environment such as materials, water and land. For example, AI used in food systems can enable better monitoring crop yields, reducing the need for chemicals and excess water through precision agriculture and minimising food waste through forecasting demand and identifying spoiled produce.
Lastly, AI can contribute to sustaining biodiversity by detecting wildlife poaching by using AI-powered image classification and object detection, and illegal logging in vulnerable forest areas by analysing audio-sensor data.
Fighting hunger and poverty
The fight against poverty and hunger is one of today’s greatest global challenges, as described by the United Nations. Poverty is caused by various factors, like lack of affordable local food, low levels of education and skills, natural disasters, and epidemics, among the most common, and there is no “silver bullet” for tackling this problem, states Elisabeth Mason, founding director of the Stanford Poverty & Technology Lab for NBS News.
But scientists have found a way to harness AI to fight this socio-economic challenge. Some of the ways researchers are employing AI to this issue is to pinpoint the regions most in need, design solutions to improve agriculture, as well as increase access to education and improve literacy.
AI is used in a myriad of use cases in healthcare. From improving diagnosis, to enhancing drug discovery and production, and virtual healthcare assistants. One remarkable example of robot-assisted surgery comes from Intuitive Surgical, a leader in the medical robotics space. Their da Vinci Surgical System with more than 5,000 surgical robots completed over one million surgical procedures including neurology, gynecology, orthopedics, urology, dental work, and hair transplants. The robots proved they can stitch wounds with greater accuracy and remove tumours with more caution which caused less damage to surrounding tissue.
Another AI application for social good in healthcare is using predictive analytics techniques for predicting the risk of cancer. PathAI has developed an AI system that helps healthcare professionals in the qualification of the PD-L1 protein, a protein that cancer cells contain in large amounts, in order to determine if a patient is more likely to develop cancer.
Amnesty International started the Troll Patrol project – an initiative that gathered human rights researchers, technical experts and thousands of online volunteers, who with the help of AI, data science and machine learning, identified and quantified online abuse against women on Twitter.
The Troll Patrol project was conducted in collaboration between Amnesty International and Element AI, a global artificial intelligence software product company. The teams revealed the shocking scale and nature of online abuse faced by women. Apart from it, the findings also provide a resource to researchers and engineers interested in exploring the potential of machine learning in content moderation.
AI and United Nations’ “Sustainable Development Goals” (SDG)
McKinsey have mapped out the United Nations’ Sustainable Development Goals (SDGs), which are among the most frequently cited societal challenges, to their AI use cases, covering all 17 of UN’s goals and supporting some aspect of each one.
The chart mainly focuses on effective management in the public and social sectors, or use cases related to disaster response and search and rescue in the crisis-response domain. Only 21 out of the total of 156 McKinsey AI use cases don’t target any of UN’s Sustainable Development Goals.
Image source: McKinsey Global Institute analysis
Moreover, they have broken down how different AI capabilities like computer vision, natural-language processing, and speech and audio processing, can be used for social good to benefit society.
Image source: McKinsey Global Institute analysis
Social justice, diversity, equity, inclusion and sustainability with AI
The Australian Red Cross is working on AI governance for sustainability because AI has a huge potential for humanity and significant risk if not properly governed, emphasises Dr. Mahendra Samarawickrama, Senior Manager – Data Science and Analytics at Australian Red Cross in his interview for Hyperight.
It is predicted that AI can contribute as much as 15.7 trillion to the world economy by 2030. At the same time, AI creates great risks for humanity such as autonomous weapons, automation-spurred job loss, socio-economic inequality, privacy violations, Deepfakes, and bias caused by data and algorithms.
Further, it has been predicted that by 2022, 85% of AI projects will fail due to bias in data, algorithms, or the teams responsible for managing them. One way of addressing issues in AI governance is to enhance AI ethics by ensuring diversity, equity, inclusion and social justice, Dr. Samarawickrama states.
Red Cross has unified its 7 fundamental principles together with the United Nations’ Sustainable Development Goals (SDGs) to drive sustainable AI for sustainability. This drive is further strengthened by their co-competencies in humanitarian initiatives, volunteering, partnership and community engagement. Together it enhances social diversity, equity and inclusion while creating an environment to cooperate for engaging in social responsibilities with the Red Cross.
Dr. Samarawickrama is going to present Red Cross’ Unified AI-Ethics Collaboration Framework for Corporate Social Responsibility (CSR) at the Data Innovation Summit. Their AI-Ethics Framework for Humanity and Sustainability enables adopting AI ethics as an important state in an organization’s AI roadmap which drives AI for serving humanity and future sustainability.
Challenges with AI for social good that need to be tackled
The growing amounts of data coupled with advanced analytics capabilities have provided us with the opportunity to solve some of the biggest social, economic and environmental challenges humanity is dealing with.
However, there are some standing challenges that must be overcome if we want to realise the potential of AI applied for social good.
Some of the most significant challenges McKinsey described are:
- Data needed for social impact is not easily accessible because much of the data essential or useful for social-good applications are in private hands or in public institutions that might not be willing to share their data.
- The expert AI talent needed to develop and train AI models is in short supply.
- ‘Last-mile’ implementation challenges for AI deployment for social good. NGOs and social-sector organizations can face technical problems with deploying and sustaining AI models that require continued access to some level of AI-related skills. Also, they may have difficulties in interpreting the results of the AI model.
On top of the above bottlenecks, Sean Lang also outlined several challenges in his Future Says… AI for Good article. Although a lot of organisations are working in this space, there is still a lack of coordination amongst these players, which leads to one-off projects, duplication of work and a drain on much-needed resources.
Additionally, most of the AI4SG initiatives are done as philanthropic pro bono projects which means companies are not compensated for their costs, not to mention generating any profit, emphasises Sean. As many of the initiatives are in the pilot or research phase, they require little investment from companies, which is not sustainable when they need to be operationalised with continuous data feeds.
Not to mention the ethical aspect of the AI being used for those in great need. Bias, privacy and safety are some of the risks that need to be mitigated when applying AI for social good. One of the most pronounced is the lack of government intervention and digital maturity when it comes to AI investment to address these social challenges, adds Sean.
Ultimately, we have in our hands the most powerful technology to solve the biggest problem posed to humanity. It’s up to public, private and NGO organisations to come together and create better collaboration in order to leverage AI4SG for a better future.