The Association for the Advancement of Artificial Intelligence (AAAI) created Artificial Intelligence for Social Good or AI4SG as an emerging topic at their annual conference in 2019. Why is it emerging? AI has gotten so much negative press. It controls what we see and possibly what we think through social media like Facebook and Instagram, it has perpetuated discrimination through poorly designed facial recognition programs and it will automate us all out of jobs, right? Perhaps. David Wong, Philosophy Professor at Duke University says, “New technology is not good or evil in and of itself. It’s all about how people choose to use it.”
Technology tells us what IS possible but not what SHOULD be done. Should nuclear weapons ever be used? Will widespread surveillance make the public safer? Should gene editing be used to eradicate congenital diseases? Science will not make these decisions, people will.
WWII let to the Universal Declaration of Human Rights which outlined the fundamental freedoms that every human should be given. The Belmont Report in 1978 outlined the ethical principles and guidelines that must be followed by anyone performing research on human subjects and had a profound impact on research methods in medicine and the social sciences. It required respect for individuals, to maximize benefits and minimize risks for research subjects and the fair treatment of subjects through reasonable and rationalized procedures.
Computer Science has not had these constraints put on it. Historically academic computer science programs taught little to no ethics. Initially computer science technology had relatively little impact on the lives of the general public. Boy, has that changed! Algorithms in medical data, social media data and financial data can have a significant impact on society and on individuals. This rapid increase in power and reach of computer technology with very few legal precedents to control the landscape means it is unregulated. As the push grows to make computer science more ethical there is increased focus on transparency and accountability.
Enter AI for Social Good or AI4SG. The AAAI established the criteria required for something to be considered AI4SG: The social impact of the problem is considered significant and not adequately addressed by the AI community, it is a novel approach and the scope and promise for social impact as well as the advancement of cutting edge AI techniques exists.
What are projects that meet the AI4SG criteria? TowardsDataScience.com has an interesting and detailed list including counting penguins from space, countering elephant poaching, substance abuse prevention, the global immunological observatory for pandemic prediction, social media analysis for mental health monitoring and perhaps the most well-known, Google’s project to decipher speech from individuals with severe speech impediments, like those with ALS, and enable them to better communicate.
Interested in being a part of the growth of AI4SC? Contact Smith Hanley Associates’ Data Science and Analytics Recruiter Nancy Darian at ndarian@smithhanley.com.
