Three Laws Blog: Law and Artificial Intelligence

Will Robots Free Us From Work? @ConversationUK

Over at the Conversation, a short discussion of the possibility that robots (of any kind) will eventually mean we can all retire with fat pensions. I have to agree that I don't think Wall-E is in our futures, for a number of reasons, including that most people have to have a reason to get out of bed in the morning, and work is a good one. Additionally, people want to be remembered, and children aren't the only way to accomplish that. Obviously one could commit crimes in order to become famous, but that kind of activity has a severe downside. What most people want, I think, is to be immortal.  Our accomplishments in life are a means to make certain the next generation, and maybe the one after that one, remembers us.

Raymond, Young, and Shackelford on Building a Better HAL 9000: Algorithms, the Market, and the Need to Prevent Engraining Bias in AI

Anjanette Raymond, Indiana University School of Business; Queen Mary University of London, School of Law, and Indiana University (Bloomington) School of Law, Emma Arrington Stone Young, Indiana University (Bloomington), and Scott Shackelford, Indiana University School of Business, Harvard Kennedy School Belfer Center for Science & International Affairs; Center for Applied Cybersecurity Research; Stanford Center for Internet and Society; Stanford Law School, are publishing Building a Better HAL 9000: Algorithms, the Market, and the Need to Prevent the Engraining of Bias in the 2017 volume of the Northwestern Journal of Technology and Intellectual Property. Here is the abstract.

As sci-fi fans will recall, the movie 2001: A Space Odyssey is focused on the interaction between humans and artificial intelligence. In the movie, HAL (Heuristically programmed Algorithmic Computer) 9000 computer is an artificial intelligence and the onboard computer on the spaceship Discovery 1. HAL 9000, more commonly called “Hal,” is capable of many functions, such as speech, facial recognition, lip reading, interpreting emotions, and expressing emotions. HAL is built into the Discovery 1 spacecraft, and is in charge of maintaining all mechanical and life support systems on board. As the movie progresses, the astronauts become concerned about HAL’s behavior and agree to disconnect him, in essence killing HAL. HAL becomes aware of the plan and seeks to stop his death as the movie plot climaxes in a conflict between intelligent machine and his human controllers. Interestingly, 2001: A Space Odyssey author Arthur C. Clark could not have been more accurate about one of the emerging conflicts to face humanity: what role does society want machines to play in coordinating and governing human activity? The debate resonates from the shared economy to the ethics of artificial intelligence. This article seeks to advance the debate about the need for data regulation that focuses on the impact of the use of the data. First, it provides a brief explanation of data analytics, algorithms, and machine learning. Second, the article explores some of the common mistakes associated with data modeling within algorithmic processes. Third, the paper explores the impact of the use of data, specifically data that is used to create a digital personhood, to inform algorithms that perform basic services. Fourth and finally, the article seeks to define an ethical decision-making model and regulatory structure for data focusing on the impact of the use of the data upon the individual and society.

Download the article from SSRN at the link.

Etzioni and Etzioni on Designing AI Systems That Obey Our Laws and Values @amitai_etzioni @etzioni

Amitai Etzioni, George Washington University, and Oren Etizioni, Allen Institute for Artificial Intelligence, have published Designing AI Systems That Obey Our Laws and Values in volume 59 of the Communications of the ACM (September 2016). Here is the abstract.

More and more machines -- from driver-less cars to weapons to search engines -- are using AI as part of their operational system. Because these systems are semi-autonomous and learning on their own from experience -- they are likely to stray from their programmers' instructions. These cars may learn to speed; AI-equipped drones hit wrong targets; and search engines discriminate. We suggest that to prevent such problems, society needs a whole new kind of AI programs, supervisory ones, that will monitor, audit, or correct the operational AI programs. In some cases they will ensure that laws are observed. In others, that the machines follow the values of their users, a special kind of AI program, the ethics bot, that will discern these values by observing the users of the machines.

Download the article from SSRN at the link.

What Happens When a Robot Decides To Click the "I Am Not a Robot" Captcha

What happens? The Empire is defeated. Of course, I'm not sure exactly why this robot clicked on "I am not a robot." Did its programming tell it to do so? Or did it actually recognize that it is a robot, but it no longer wants to be a robot, thus overcoming its programming and achieving its (non) robot destiny? Too deep for me. More here.

Very Inappropriate Robot Gavels @igavels

Bloomberg Business News features this piece about robots and their use in domestic relations law. The image would be fine if it pictured a U.S. judge--they use gavels. But it shows a robot in a wig, and U.K. and Canadian judges, for example, don wigs, when they're presiding in court. And they don't use gavels. Nor do Australian judges. 

I'm suggesting a Fifth Law of Robotics: A robot must not use gavels inappropriately. 

Human Programmers and AI Prejudice

From Scientific American, how AI learns bias.  Hmm. Not looking good for IDIC. 

The Ethics of Artificial Intelligence

From NPR, a consideration of the issue, beginning with Isaac Asimow's Three Laws of Robotics. Contemporary thinkers note that things have moved along since then, beginning with the introduction of self-driving cars. 

Now Interviewing--Robots

What jobs are robots likely to take over? Check out this infographic from Euroffice, based on sources from the last few years.

Google and the Developing Laws of Robotics

According to this article on Extreme Tech, Google is working to develop rules that will direct the development of AI along paths that guarantee that robots, androids, and other artificial intelligence progress in ways that humans and not AI determine, which should address those doomsday scenarios SF screenwriters and novelists keep presenting us.  That is, they're finally beginning to grapple with the issue of how AI thinks, not what it thinks. Philip K. Dick, begone!

Mr. Robot Returns

A recap of the first season of Mr. Robot, courtesy of The Hollywood Reporter, just in time for the season 2 premiere. And commentary from Slate on the show, here and here.