{"id":7064,"date":"2019-04-17T14:01:56","date_gmt":"2019-04-17T21:01:56","guid":{"rendered":"https:\/\/www.springboard.com\/?p=7064"},"modified":"2023-10-10T00:41:52","modified_gmt":"2023-10-10T07:41:52","slug":"ai-for-good","status":"publish","type":"post","link":"https:\/\/www.springboard.com\/blog\/data-science\/ai-for-good\/","title":{"rendered":"AI for Social Good: 7 Inspiring Examples"},"content":{"rendered":"\n

Over the past decade, rapid advancements have made it possible for AI systems to do things we once only dreamed about. However, much of the hype around AI and machine learning tends to focus on its potential for business, productivity, and profits. Perhaps there should be more spotlight on how we can use AI for good. <\/span><\/p>\n\n\n\n

AI has the power to tackle many of the biggest problems on the planet and could make a huge impact on sustainability, our environment, and even humanity itself. As you\u2019ll see from the real-life examples in this post, robots, and humans are already showing they can be an incredible team.<\/span><\/p>\n\n\n\n

7 Examples of the World Using AI for Good<\/b><\/h2>\n\n\n\n

You may wonder how AI and sustainability fit together, or how machine learning can help the environment. <\/span><\/p>\n\n\n\n

The immense potential of these technologies is still being discovered. For anybody who wants to <\/span>start a career in AI<\/span><\/a>, machine learning, or data science<\/a>, you may be pleasantly surprised to learn you can do a lot more than analyze consumer behaviors and company profit margins. <\/span><\/p>\n\n\n\n

Here are seven examples of how our world is using AI for good today.<\/span><\/p>\n\n\n\n

Water Management<\/span><\/h3>\n\n\n\n
\"Water<\/figure>\n\n\n\n

(Source<\/span><\/a>)<\/p>\n\n\n\n

As a large region with hot, dry summers, Southern California needs to take water use seriously. Recent years have brought long droughts, which can cause havoc for residents. <\/span><\/p>\n\n\n\n

When the Moulton Niguel Water District (MNWD) was expecting the worst drought in 500 years, the district turned to <\/span>DataKind<\/span><\/a>, an organization that focuses on using <\/span>\u201cdata science in the service of humanity.\u201d<\/span><\/i><\/p>\n\n\n\n

DataKind collected flow data from the microzone level, as well as data on water usage, weather, and the population, to develop a <\/span>forecasting model<\/span><\/a>. This data infrastructure allowed MNWD to accurately predict water demand, which helped save the district around $5 million that would otherwise have been invested in capital storage and importing potable water.<\/span><\/p>\n\n\n\n

In the graph above, the blue line indicates the forecasted water demand, and it is surrounded by a confidence band. Whenever a drought strikes, the demand line may reach the red line. Once it gets close, the AI system will trigger a warning to an MNWD analyst. At that point, the MNWD can communicate with their largest water consumers and take precautions to control water use in the area.<\/span><\/p>\n\n\n\n

Global Warming<\/span><\/h3>\n\n\n\n
\"AI<\/figure>\n\n\n\n

(Source<\/span><\/a>)<\/p>\n\n\n\n

Bill Gates has been a champion of environmental action for many years, so it comes as little surprise to see Microsoft is investing heavily in using AI for social good and sustainability. The company that Gates founded has spent over $50 million on its <\/span>AI for Earth<\/span><\/a> program, which creates and tests innovative AI applications.<\/span><\/p>\n\n\n\n

Climate change is an ongoing challenge, especially when natural disasters ravage our jungles and forests, \u201cthe lungs of our planet.\u201d<\/span><\/p>\n\n\n\n

Whenever the distribution of different tree species is impacted, vegetation decomposition rates change and more CO2 enters our atmosphere. Over time, trees will grow back smaller and more storms will exacerbate the cycle, resulting in a CO2-heavy environment. This speeds up global warming.<\/span><\/p>\n\n\n\n

In the wake of <\/span>Hurricane Maria<\/span><\/a>, the AI for Earth program awarded a grant to Maria Uriarte, an ecology professor from Columbia University. Uriarte\u2019s team included a statistics professor, and together with their Microsoft grant, they set out to Puerto Rico to study the impact of the hurricane on the El Yunque National Forest. <\/span><\/p>\n\n\n\n

By using AI and data science, they can analyze how trees have been disrupted in the forest. Without AI, such a study wouldn\u2019t be possible. If we can develop a better understanding of how our forests change in the wake of hurricanes, we can look for measures to combat climate change.<\/span><\/p>\n\n\n\n

Renewable Energy<\/span><\/h3>\n\n\n\n
\"wind<\/figure>\n\n\n\n

(Source<\/span><\/a>)<\/p>\n\n\n\n

It\u2019s been said that \u201cdata is the new oil.\u201d It\u2019s just as well, because we\u2019re burning through our oil pretty fast, and may not have any left in about <\/span>50 years<\/span><\/a>. So, how can AI and sustainability come together to help us develop our renewable energy sources?  <\/span><\/p>\n\n\n\n

As a carbon-free form of electricity, wind power is on the rise. It\u2019s getting cheaper, and it is gradually being embraced. The one sticking point, however, is the unpredictable nature of the wind. Without a wind source, it\u2019s hard to run a turbine consistently at the level a city would demand. <\/span><\/p>\n\n\n\n

Google<\/span><\/a> has sought to find the answer to this problem by teaming up with DeepMind. Together, they conducted testing of machine learning algorithms on 700 megawatts of wind power capacity, and afterward, Google claimed the technology added 20% to the value of the wind power.<\/span><\/p>\n\n\n\n

This significant gain has been made possible by a neural network that comprises turbine data and weather forecasts. The DeepMind AI system can use this data to accurately predict the output up to <\/span>36 hours before<\/span><\/i> the wind power needs to be generated. This system can help the world optimize its electricity use on large power grids, effectively conserving energy in the long run.<\/span><\/p>\n\n\n\n

Earthquakes<\/span><\/h3>\n\n\n\n
\"ConvNetQuake\"<\/figure>\n\n\n\n

(Source<\/span><\/a>)<\/p>\n\n\n\n

If you\u2019ve ever experienced an earthquake, you\u2019ll know how frightening it can be. Typically, we only feel quakes that reach <\/span>2.5 on the Richter scale<\/span><\/a>, but thousands happen every day without us knowing. <\/span><\/p>\n\n\n\n

In the past decade, major quakes in Haiti, Japan, and New Zealand totaled entire cities and claimed thousands of human lives. Without adequate methods of predicting events or preparing for major earthquakes or aftershocks, there was no time for people to get to safety.<\/span><\/p>\n\n\n\n

Now, as artificial intelligence and big data have become more sophisticated, we have both the technology and the necessary data scientists to study earthquakes at a deeper level. <\/span><\/p>\n\n\n\n

Thibaut Perol is a deep learning scientist who is leading an AI project called <\/span>ConvNetQuake<\/span><\/a>. This system utilizes a machine learning algorithm that analyzes seismograms to determine whether the seismic activity is an earthquake or simply low-level noise. <\/span><\/p>\n\n\n\n

Testing in Oklahoma was a success, with <\/span>17 times as many earthquakes<\/span><\/a> detected compared to traditional measures. This makes ConvNetQuake the first neural network in the world that can detect earthquakes, even low-level ones that traditional practices fail to pick up. <\/span><\/p>\n\n\n\n

With greater investment and development, it may be feasible that the world will have a reliable AI system that can warn us ahead of major quakes, which could save thousands of lives.<\/span><\/p>\n\n\n\n

Poaching and Illegal Fishing<\/span><\/h3>\n\n\n\n
\"PAWS<\/figure>\n\n\n\n

(Source<\/span><\/a>)<\/p>\n\n\n\n

Some of the world\u2019s animals are on the verge of extinction because of poaching, with around <\/span>100 elephants<\/span><\/a> killed every day for ivory, meat, and body parts. Rhinos and tigers are also under severe threat, as poachers relentlessly hunt these endangered species for their horns or fur. <\/span><\/p>\n\n\n\n

Conservation agencies face a huge challenge in patrolling large natural reserves. Equipped with limited resources, their personnel are fighting a losing battle. However, help may be on the horizon.<\/span><\/p>\n\n\n\n

Protection Assistant for Wildlife Security, better known as <\/span>PAWS<\/span><\/a>, is an AI-led system that evaluates information about previous patrols and poaching activity, then generates logical patrol routes for conservation agencies to scout. Once the patrol units complete the route, they can feed more data back into PAWS, such as:<\/span><\/p>\n\n\n\n