AI For Good: Using Generative AI As A Tool To Better The World

When it comes to AI and its uses, opinions are generally pretty polarized. Many people believe that it is a potentially dangerous or unnecessary invention while others believe that it is a very valuable development that has few to none faults or potential negative consequences. While both opinions have their place, often this polarization is developed from misconceptions about what AI is and many of its more common uses, specifically for how it can be used for good.

When many people think of AI, they think of generative AI programs like Chat GPT which respond to specific prompts given by humans, something that people often use in a more casual way, such as to answer a question. This one perception of AI looks over the many other uses of AI technology, namely, how it can often be used in a scientific setting to help people and to develop new programs for the benefit of society.

Mapping Potential Flooding

Researchers at MIT were able to develop an AI program that is able to generate possible satellite imagery of flooding after a natural disaster, a tool that could help civilians decide whether to evacuate or not and possibly save countless lives.

This program combines a generative AI model and a physics-based flood model to provide a realistic depiction of flooding that could potentially occur after a hurricane. Combining generative AI with a physics-based model allows for a more realistic depiction of this potential flooding, as using only generative AI caused researchers to get results that were not possible.

Since this development is so new, it needs to be trained on many more depictions of natural disasters before it can be used as a reliable source for potential flooding in certain regions. Once this technology is ready for widespread-use, it can serve as an incentive for many people to evacuate from these natural disasters because they will be able to visualize the flooding that could occur. This could save so many lives through the element of visualization encouraging people to evacuate.

Using generative AI to map out potential flooding from hurricanes is a small part of efforts to map out various future climate possibilities, such as reforestation or ice melt.

Lowering Disease Risk

Researchers at Harvard University have also developed ways of using AI to better society, specifically in the realm of lowering HIV risks. Studied have shown that using peer leaders within groups at risk for HIV (for example: homeless youth) to promote healthy habits like condom usage and getting tested for STIs can help lower risk of HIV. However, this is specifically tied to choosing the right peer leader.

Researchers were able to use an AI algorithm to select the most effective peer leader among these at risk groups, which was directly connected to the lowering of the risk of HIV. This algorithm took the participants of this study (which included 700 members in total) and used social mapping to find who had the most connections to many different social groups rather than simply the most popular participants.

This method of social mapping using generative AI quickly lowered risk of behaviors that could lead to HIV, as compared to the much slower impact of only choosing the most popular member of a group.

This research could be a big step in being able to use social networks in order to quickly spread information about health and wellness, something that could help many different communities learn about public health issues more efficiently and effectively.