AI Matters: our blog

News from USACM

Next week the  USACM Council will be holding its annual in-person meeting in Washington, beginning with a reception Wednesday, March 21st from 5 to 7 at the Georgetown home of Law Committee Chair Andy Grosso. We cordially invite DC-area USACM members to join us. If you plan to attend, please RSVP to Adam Eisgrau <eisgrau@HQ.ACM.ORG>, who will provide further details.

Statement of the European Group on Ethics in Science and New Technologies on “Artificial Intelligence, Robotics and ‘Autonomous’ Systems,” published March 9:
http://ec.europa.eu/research/ege/pdf/ege_ai_statement_2018.pdf
The statement calls for the EC to “launch a process that paves the way towards a common, internationally recognized ethical and legal framework for the design, production, use and governance of artificial intelligence, robotics, and ‘autonomous’ systems.”

President Donald Trump today tapped Obama-era deputy U.S. CTO Ed Felten to serve on the Privacy and Civil Liberties Oversight Board (https://www.pclob.gov/).

ACM SIGAI Learning Webinar “Advances in Socio-Behavioral Computing”

This live presentation was given on Thursday, March 15  by Tomek Strzalkowski, Director of the Institute for Informatics, Logics, and Security Studies and Professor at SUNY Albany. Plamen Petrov, Director of Cognitive Technology at KPMG LLP and Industry Liaison Officer of ACM SIGAI, and Rose Paradis, Data Scientist at Leidos Health and Life Sciences and SIGAI Secretary/Treasurer, moderated the questions and answers session.

Slides are available here.

This talk presented ongoing research on computational modeling and understanding of social, behavioral, and cultural phenomena in multi-party interactions. They discussed how various linguistic cues reveal the social dynamics in group interactions, based on a series of experiments conducted in virtual on-line chat rooms, and then showed that these dynamics generalize to other forms of communication including traditional face-to-face discourse as well as the large scale online interaction via social media. They also showed how language compensates for the reduced cue environment in which online interactions take place.

They described a two-tier analytic approach for detecting and classifying certain sociolinguistic behaviors exhibited by discourse participants, including topic control, task control, disagreement, and involvement, that serve as intermediate models from which presence the higher level social roles and states such as leadership and group cohesion may be inferred. The results of an initial phase of the work used a system of sociolinguistic tools called DSARMD (Detecting Social Actions and Roles in Multiparty Dialogue).

Several extensions of the basic DSARMD model move beyond recognition and understanding of social dynamics and attempt to quantify and measure the effects that sociolinguistic behaviors by individuals and groups have on other discourse participants. Potentially, autonomous artificial agents could be constructed capable of exerting influence and manipulating human behavior in certain situations. Such extended capabilities could possibly be deployed to increase accuracy of predicting online information cascades, persuasion campaigns, and even defend against certain forms of social engineering attacks.

The model and tools presented in the Webinar are interesting to consider in the detection and assessment of algorithmic bias.

The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18) was on  February 2–7, 2018, at the Hilton New Orleans Riverside.  The AAAI/ACM Conference on AI, Ethics, and Society (AIES) was held at the beginning of AAAI-18. Developers and participants included members of SIGAI and USACM.

The AIES conference description follows: “As AI is becoming more pervasive in our life, its impact on society is more significant and concerns and issues are raised regarding aspects such as value alignment, data handling and bias, regulations, and workforce displacement. Only a multi-disciplinary and multi-stakeholder effort can find the best ways to address these concerns, including experts of various disciplines, such as
ethics, philosophy, economics, sociology, psychology, law, history, and politics. In order to address these issues in a scientific context, AAAI
and  ACM have joined forces to start this new conference.”

The full schedule for the AIES 2018 Conference is available at www.aies-conference.com. A panel relevant to our policy blog discussions “Prioritizing Ethical Considerations in Intelligent and Autonomous Systems – Who Sets the Standards?” was designed by our IEEE/ACM committee and will be covered in a future post.

In the next few blog posts, we will present information and generate discussion on policy issues at the intersection of AI, the future of the workforce, and educational systems. Because AI technology and applications are developing at such a rapid pace, current policies will likely not be able to impact sufficiently the workforce needs even in 2024 — the time frame for middle school students to prepare for low skill jobs and for beginning college students to prepare for higher skilled work. Transparency in educational policies requires goal setting based on better data and insights into emerging technologies, likely changes in the labor market, and corresponding challenges to our educational systems. The topics and resources below will be the focus of future AI Policy posts.

Technology

IBM’s Jim Spohrer has an outstanding set of slides “A Look Toward the Future”, incorporating his rich experience and current work on anticipated impacts of new technology with milestones every ten years through 2045. Radical developments in technology would challenge public policy in ways that are difficult to imagine, but current policymakers and the AI community need to try. Currently, AI systems are superior to human capabilities in calculating and game playing, and near human level performance for data-driven speech and image recognition and for driverless vehicles. By 2024, large advances are likely in video understanding, episodic memory, and reasoning.

The roles of future workers will involve increasing collaboration with AI systems in the government and public sector, particularly with autonomous systems but also in traditional areas of healthcare and education. Advances in human-technology collaboration also lead to  issues relevant to public policy, including privacy and algorithmic transparency. The increasing mix of AI with humans in ubiquitous public and private systems puts a new emphasis on education for understanding and anticipating challenges in communication and collaboration.

Workforce

Patterns for the future workforce in the age of autonomous systems and cognitive assistance are emerging. Please take a look at Andrew McAfee’s presentation at the recent Computing Research Summit. Also, see the latest McKinsey Report — Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation. Among other things, this quote from page 20 catches attention: “Automation represents both hope and challenge. The global economy needs the boost to productivity and growth that it will bring, especially at a time when aging populations are acting as a drag on GDP growth. Machines can take on work that is routine, dangerous, or dirty, and may allow us all to use our intrinsically human talents more fully. But to capture these benefits, societies will need to prepare for complex workforce transitions ahead. For policy makers, business leaders, and individual workers the world over, the task at hand is to prepare for a more automated future by emphasizing new skills, scaling up training, especially for midcareer workers, and ensuring robust economic growth.”

Education for the Future

An article in Education Week “The Future of Work Is Uncertain, Schools Should Worry Now” addresses the issue of automation and artificial intelligence disrupting the labor market and what K-12 educators and policymakers need to know. A recent report by the Bureau of Labor Statistics “STEM Occupations: Past, Present, And Future” is consistent with the idea that even in STEM professions workforce needs will be less at programming levels and more in ways to collaborate with cognitive assistance systems and in human-computer teams. Demands for STEM professionals will be for verifying, interpreting, and acting on machine outputs; designing and assembling larger systems with robotic and cognitive components; and dealing with ethics issues such as bias in systems and algorithmic transparency.