Artificial Intelligence Roadmap

< Back to AI Roadmap Landing Page

2. Major Societal Drivers for Future Artificial Intelligence Research

AI has permeated our lives and has become an engine for innovation across all sectors of society. Government investments can have a profound impact and transform society for the betterment of all citizens. Below we describe six societal areas that will be transformed by AI. These are meant to be examples of AI’s potential impact to society, but in no way are they an exhaustive list. Detailed discussions for each of the societal drivers are included in the workshop reports through a series of vignettes that describe how a person or organization could interact with AI systems or would benefit from AI technologies, and discuss the capabilities that would be required and that motivate the AI research in this Roadmap. By looking at and working on the research challenges grounded in AI for social good, we will not only be changing society, but we can motivate K-12 students to study AI by presenting exciting problems to be worked on. These same societally relevant problems will help to engage women and other underrepresented populations in AI research, creating a more diverse workforce to better tackle the problems. In highlighting AI for social good, we will dispel misconceptions about AI while solving real-world, tangible problems.

2.1 Enhancing Healthcare and Quality of Life

Although the potential benefits for AI in healthcare have been demonstrated, this technology is largely untapped in clinical settings. AI applications can now diagnose some skin cancers more accurately than a board-certified dermatologist, and do so faster and more efficiently.[1]

Figure 2. Identifying Societal Drivers

Figure 2. Identifying Societal Drivers

AI techniques already developed can examine high-risk spots on mammograms and provide advice on whether a biopsy is necessary, potentially reducing the number of biopsies by 30 percent.[2] In the near term, chronic health conditions like diabetes, cancer, and heart and neurological diseases are likely to benefit most from new applications of AI, according to a survey of healthcare professionals.[3] In addition to AI’s ability to aid in the medical diagnosis process, the deployment of AI applications in hospital workflows could further enhance care delivery. An analysis by Frost & Sullivan suggests such cross-cutting uses of AI “have the potential to improve patient outcomes by 30 to 40 percent, while cutting treatment costs by as much as 50 percent.”[4] The impacts of AI enabled technologies over the longer term are likely to be even more profound.

2.2 Lifelong Education and Training

The US education system currently makes use of computing technologies, some of which are enhanced by AI, throughout the education environment, but there is still great room for improvement of these technologies, and even greater opportunity for full adoption to enhance our education system. In K-12 classrooms, students use adaptive reading and math software from a variety of vendors to receive personalized curricula tailored to their own progress. Content providers use machine learning to determine what material works well in each area. Teachers make use of feedback and scoring systems to help grade assignments, guard against plagiarism, and assess student progress. Augmented reality and virtual reality are becoming useful teaching and training tools, and AI systems that understand gesture and voice recognition are delivering more effective simulations. Assistive technologies powered by AI technologies are helping special-needs students recognize voices and text, thereby easing their communication. Even well outside the classroom, technology is changing how education is delivered in the US Student transportation services are increasingly optimized using AI, as is staff scheduling and substitute management. Grounds and facilities management, including smart building-management software and intelligent security products, is aided by AI. The market for AI technologies in education and training in the US, including higher-education, K-12 education, and corporate training, was $400 million in 2017, and is forecast to grow more than 45 percent per year to about $5.4 billion by 2024.[5] This Roadmap envisions integrated interactive AI systems that will expand education and career opportunities across broader segments of our society and level the playing field for students facing socio-economic, health, and disability challenges in our nation’s schools.

2.3 Reinvent Business Innovation and Competitiveness

AI is helping drive US business innovation and competitiveness. Indeed, it is arguable that US industries’ embrace of the AI technologies birthed from fundamental work in our research labs and institutions, notably machine learning, has provided a key competitive advantage to US companies in maintaining a leadership role in the global economy. By some estimates, AI contributed $2 trillion to global GDP in 2018 and is expected to be $15 trillion by 2030.[6] Business leaders worldwide see AI technologies as increasingly crucial to their competitiveness. When surveyed, executives at companies deploying AI cite its capability for generating new revenue, helping retain existing customers and acquire new ones, and providing a competitive differentiation with other companies in their sector as their top reasons for using the technology.[7] However, because of the limitations of the current state of the art, these deployments tend to be somewhat narrowly focused, applied to particular problems and constrained by the inability of current AI systems to integrate intelligence from a wide range of sources and contexts. The Roadmap describes a broader set of capabilities for business innovation and competitiveness, from developing broad ranging recommender systems to integrating robotic and AI systems into standard practices.

2.4 Accelerate Scientific Discovery and Technological Innovation

Some areas of modern science are benefiting from the use of AI technologies. The oceans of experimental and observational data produced by today’s scientific infrastructure—e.g., space telescopes, super-colliders, sequencing equipment, medical imaging—are simply too vast for humanity to wade through unaided. Machine learning algorithms and other AI technologies help us make sense of the chaos by noting anomalies and detecting patterns that would otherwise go unnoticed.[8] Yet we are still only working in narrow models. The Roadmap envisions new hybrid modeling approaches that could be facilitated via advanced AI learning research. Advancements in machine learning techniques will enable techniques to process multimodal, multi-scale data, handle heterogeneity in space and time, and accurately quantify uncertainty in the results. The continued improvement of AI systems will impact the pillars of science, from data collection to experiment to discovery.

2.5 Social Justice and Policy

There is no greater opportunity to enhance quality of life than by increasing the equity, effectiveness, and efficiency of public services provided to citizens. The opportunity for AI to have an impact in the social policy and justice space is tremendous. McKinsey has recently done an analysis of 160 AI social-impact use cases and determined that existing capabilities could contribute to tackling cases across all 17 of the United Nations sustainable-development goals.[9] One of the goals is to take immediate and effective measures to eradicate forced labor, end modern slavery and human trafficking, and secure the prohibition and elimination of the worst forms of child labor, including recruitment and use of child soldiers, and by 2025 end child labor in all forms.[10] Methods exist today to identify trafficking or forced labor, and methods exist to help people once they are removed from their engagement in trafficking situations, but across the globe, the translation from identification to removal is often fraught with missed opportunities. The Roadmap envisions a future where meaningful interactions between people and systems will enable a complete path from forced laborer to survivor.

2.6 Transform National Defense and Security

Artificial Intelligence is already an essential part of cyber security, providing commercial and military cyber defenders with comprehensive security monitoring, threat detection, and actionable insight. At the same time, AI provides adversaries with unprecedented ways to understand a target’s vulnerabilities and vector attacks accordingly. Spearphishing attacks are one highly effective means for an adversary to gain access to a target machine, but they are also labor-intensive in that they require research on the subject to determine what content would be most likely to motivate a click. A 2016 experiment demonstrated that AI-based spearphishing attacks on Twitter could garner almost the same response rate as a human-led attack could (34 percent vs. 38 percent), while targeting almost eight times as many victims in the same time period.[11] Defending networks from an intruder using compromised legitimate network credentials, perhaps gleaned by spearphishing, requires deep learning to analyze the user’s behavior over a series of actions.[12] Future approaches to cybersecurity and defense will benefit from the powerful capabilities provided by AI systems. This report envisions how future AI systems can aid in responses to other types of threats as well, including natural disasters.

 

 

 

[1]  Michael J. Rigby, “Ethical Dimensions of Using Artificial Intelligence in Health Care,” AMA Journal of Ethics, vol. 21, no. 2 (Feb. 2019): 121.

[2]
Richard Harris, reporter, “Training A Computer To Read Mammograms As Well As A Doctor,” National Public Radio, April 1, 2019. https://www.npr.org/sections/health-shots/2019/04/01/707675965/training-a-computer-to-read-mammograms-as-well-as-a-doctor and https://pubs.rsna.org/doi/10.1148/radiol.2017170549

[3]
Tom Sullivan, “3 charts show where artificial intelligence is making an impact in healthcare right now,” Healthcare IT News, Dec. 21, 2018.
https://www.healthcareitnews.com/news/3-charts-show-where-artificial-intelligence-making-impact-healthcare-right-now

[4]
Frost & Sullivan (press release), From $600 M to $6 Billion, Artificial Intelligence Systems Poised for Dramatic Market Expansion in Healthcare, Jan. 5, 2016. https://ww2.frost.com/news/press-releases/600-m-6-billion-artificial-intelligence-systems-poised-dramatic-market-expansion-healthcare

[5]
“Artificial Intelligence (AI) in Education Market,” Global Market Insights, Jan. 2018. https://www.gminsights.com/industry-analysis/artificial-intelligence-ai-in-education-market

[6]
Frank Holmes, “AI Will Add $15 Trillion to the World Economy by 2030,” Forbes, Feb. 25, 2019. https://www.forbes.com/sites/greatspeculations/2019/02/25/ai-will-add-15-trillion-to-the-world-economy-by-2030/#795251a71852

[7]
Warren Knight, “Is Artificial Intelligence the Future of Business?” Business2Community, Nov. 9, 2018. https://www.business2community.com/business-innovation/is-artificial-intelligence-the-future-of-business-02137459

[8]
Dan Falk, “How Artificial Intelligence Is Changing Science,” Quanta Magazine, March 11, 2019. https://www.quantamagazine.org/how-artificial-intelligence-is-changing-science-20190311/

[9]
Michael Chu, Martin Harrysson, James Manyika, Roger Roberts, Rita Chung, Peter Nel, and Ashley van Heteren, “Applying Artificial Intelligence for Social Good.” McKinsey Global Institute discussion paper, November 2018. https://www.mckinsey.com/featured-insights/artificial-intelligence/applying-artificial-intelligence-for-social-good

[10]
Sustainable Development Solutions Network, “Target 8.7.” Indicators and a Monitoring Framework for Sustainable Development Goals: Launching a Data Revolution for the SDGs. May 15, 2015 http://indicators.report/targets/8-7/

[11]
Artificial Intelligence for Cyber Security, The Workshops of the Thirty-First AAAI Conference on Artificial Intelligence: Technical Reports WS-17-01 — WS-17-15. Palo Alto, CA: The AAAI Press, 2017. https://aaai.org/Library/Workshops/ws17-04.php

[12] ibid

 

 


Full Report

Table of Contents:

 Title Page, Executive Summary, and Table of Contents (7 pages)

1. Introduction (5 pages)

2. Major Societal Drivers for Future Artificial Intelligence Research (4 pages)

3. Overview of Core Technical Areas of AI Research Roadmap: Workshop Reports

  1. Workshop I: A Research Roadmap for Integrated Intelligence (19 pages)
  2. Workshop II: A Research Roadmap for Meaningful Interaction (27 pages)
  3. Workshop III: A Research Roadmap for Self-Aware Learning (26 pages)

4. Major Findings (4 pages)

5. Recommendations (22 pages)

6. Conclusions (1 page)

7. Appendices (participants and contributors) (2 pages)

< Back to AI Roadmap Landing Page