A. The use of AI for public good
(RFI question 2)

Last updated July 28, 2016

For decades, we have been stockpiling digital information. We have digitized the history of the world’s literature and all of its medical journals. We track and store the movements of automobiles, trains, planes and mobile phones. And we are privy to the real-time sentiments of billions of people through social media. It is not unreasonable to expect that within this rapidly growing body of digital information lies the secrets to defeating cancer, reversing climate change, or managing the complexity of the global economy. We believe that many of the ambiguities and inefficiencies of the critical systems that facilitate life on this planet can be eliminated. And we believe that AI systems are the tools that will help us accomplish these ambitious goals.

Our society as a whole is paying a huge price for not knowing how to use our best insights when and where it matters. This happens because we simply do not have time to absorb all of the information being produced relevant to the tasks and decisions at hand. Data may be “the new natural resource”, but it is becoming too abundant to deploy. AI tools, and in particular cognitive tools, with their ability to ingest the enormous explosion of information in multiple forms including text, sounds, and images are a powerful tool to help people find insights and make better decisions. As the economy shifts to make use of even more data and information, cognitive systems tailored to unique circumstances and context will assist decision makers. Cognitive systems can be applied to growing array of problems to serve the public good and benefit society. This section considers examples of cognitive systems in health care, law enforcement, social services, finance, and education.

Healthcare: IBM’s Watson Oncology Advisor, an AI tool specifically optimized for cancer care, ingests the patients’ electronic medical history, pulling out pertinent clinical information to the immediate case, performs cohort analysis to find the micro-segmentation of other similar patients, evaluates the currently accepted standard of care practices as written, uses clinical expertise as taught by the top clinicians in the country to consider the available treatment options, rank-ordered by relevance, risk and preference, and reinforced with the most relevant clinical literature – all in support of helping clinicians decide on the best, evidence-based treatments for their cancer patients. The “democratization” of medical expertise can help reduce incidents of misdiagnosis and mistreatment, improving outcomes, and reducing the overall cost of medical care. Other work is now under way to adapt this system for diabetes care, and it has the potential to bring similar benefits to many other chronic diseases. Other AI systems are now able to improve the process of identifying the causal relationships between genes and disease, to help evaluate eligibility for clinical trials, and to accelerate the drug discovery process; and IBM is currently working on an AI tool that will interpret scans and x-Rays.

Social services: A team at the University of Texas, Austin, as part of their Senior level course in Watson developed a proposal for an app that would supplement ‘211’ call centers in answering questions about social services – ranging from things like, “Where can I find a shelter for tonight?” and “Can I get help paying my bills?” through to “How can I get help buying medicine for my children?” (See the video.) The team partnered with the United Way and, upon graduation, created a start-up company.

Education: Student engagement is key, especially for early learners. Cognitive systems are being used today in a toy Dinosaur called a CogniToy built by Element Path to interact with children by answering their questions, telling them riddles and stories, and performing basic educational exercises with the child – tailored to that child’s specific learning speed and strengths. (See video). Sesame Street is now exploring other ways of applying cognitive systems to improve the educational experience for students and teachers by monitoring various behavioral patterns in students to identify appropriate courses and educational materials to improve academic performance.

Finance: By lowering the cost of asking basic financial questions, cognitive systems promise to create more financial inclusion by lowering the wealth threshold that people must cross to access financial advice. An advisor can conceivably increase their client workload by one or two orders of magnitude, by offloading many of more mundane and trivial information retrieval tasks they perform, and in doing so open them up to serving many people that otherwise could not afford that advice.

Transportation: As widely reported, AI systems are rapidly moving towards enabling self-driving vehicles. These promise to dramatically reduce road congestion and thus pollution, freight costs and wasted time; and they will make roads far safer than human drivers will ever be able to. In addition, they promise to enable new business models offering different modes of car ownership based on vehicle sharing and subscription – once a car has ferried one “subscriber” to his or her destination, it can then rapidly move on to the next.

Law enforcement: IBM’s Watson Discovery Advisor helps investigators find associations and relationships whose signals may be too weak to recognize against a background of overwhelming information noise: these come close in some cases, to predicting the location of crimes and their likely perpetrators. AI systems are also assisting crime scene investigators and analysts who study criminal behavior, and to help answer critical questions in a line of inquiry for homeland protection.

Environment: AI systems can help construct environmental models for accurate prediction of pollutants, helping decision makers manage pollutant producers and carbon footprints. They will also be able to help in the preservation of important ecosystem services, for example where wetlands or forests protect a city from flash flooding.

Infrastructure: AI systems can assist with prediction of demand, supply, and use of infrastructure; planning and execution of projects; as well as prediction-based maintenance of built infrastructure. In 2013, McKinsey identified how $1 trillion per year could be saved from the bill to address the global infrastructure deficit from better project selection, delivery improvements and maximizing the capacity of existing infrastructures: it is reasonable to project that AI, especially when linked to the Internet of Things (IOT) will play a major role in these savings.

Back to summary.