Politics and The Tools of Artificial Intelligence (page 2 of 2)
By Denny Rock

Minsky to Mills

Marvin Minsky proposed frames as data structures for representing knowledge and expectations, which would let a computer system impose coherence on incoming information. Minsky's paper, "A Framework for Representing Knowledge," was influential among AI researchers and inspired the development of many high-level knowledge-representation languages. Representation tools such as inheritance, demons, default values, and perspectives led to procedures that make assumptions, tell what is relevant, and look for information. P. Winston used the example of news to describe frames, observing that news is an easy domain for frame finding and instantiation.

Meanwhile, political scientists have used the term "frame" for many years in reference to political function and content, void of computer representations. The idea of a political frame dates back to P. Converse's 1964 theory of "mass belief systems," H. Lasswell's 1941 study of "attention frames" in propaganda, W. Lippman's 1922 work on "public opinion," and J. S. Mill's 1861 theories on "minds of higher grade" and "democracy as government by discussion" to serve the discovery of truth and to cultivate intelligent individuals."

J. Farr defined a political frame as "a central organizing idea or story line that provides meaning to an unfolding strip of events, weaving a connection among them. The frame suggests what the controversy is about, the essence of the issue. Frames consist of metaphors, exemplars, catchphrases, depictions, and visual images; they often include a rudimentary causal analysis and appeals to honored principles. We believe that frames lead a double life, that they are structures of the mind that impose order and meaning on the problems of society and that they are interpretative structures embedded in political discourse."

Consider the following three examples of frames used in reference to the attitudes expressed in court decisions, speeches by prominent public officials, and opinions in news sources and political journals. The first example concerns civil rights. Supporters of affirmative action typically have referred to the need for "remedial action," while opponents have argued that affirmative action constitutes "unfair advantage" or "reverse discrimination." One debate now before the Supreme Court over the shapes of congressional districts of "majority minority" constituency has frames of "tyranny of the majority" and "racial gerrymandering."

A second example concerns the events surrounding the 1991 Persian Gulf conflict. The U.S. public was ignorant of general knowledge about the Persian Gulf region and many specific details. Proper framing could have alerted the public to the increasing dangers before 1991, but fragmented stories were left to stand on their own. If the public had been informed properly, the armed conflict still might have been the same. The result, according to Bennett, was unchallenged manipulation of news before the conflict and a state of political impasse afterwards. MIT linguistics professor Noam Chomsky calls this manipulation "the manufacturing of consent."

A third example concerns the debate over President Clinton's 1994 State of the Union address frame of "three strikes and you're out" for lifetime imprisonment of repeat criminals. While the rhetoric is wildly popular, it almost certainly will not do what people want reduce their risk of being victims of random violence, according to Jerome Skolnick, president of the Society of Criminology (University of California at Berkeley). The fact is that violent crimes are committed disproportionately by young men aged 13 23; their criminal activity diminishes sharply as young offenders enter their 30s. Skolnick suggests that this "bumper sticker" solution will result in a very expensive prison-building program, while not concentrating on the young who are entering criminal careers.

The creation and use of political frames is more than just vocabulary; it involves symbols, emotions, and notions of justice. Information becomes important when it is relevant to a common purpose, which is built on a set of values and relationships. Facts are assembled and interpreted differently, depending on the frames and broader system of explanation. This development, in turn, contributes to the way citizens participate in debates and the formation of public policy.

Farr warns that elite frames can serve manipulative interests of political elites, in which leaders do most of the conversing and democratic discussion is reduced to campaigning for elections and the casting of votes. He comments that "many political frames are more nationalistic, patriotic, heroic, theistic, familistic, or individualistic than they are democratic." Farr suggests that even the term "democracy" has been "introduced trivially, incoherently, or manipulatively into all sorts of domestic debates, military interventions, consumer advertisements, and television specials."

Creation of computer applications to enhance democratic discussion eventually will use these frame concepts. The old politics often depicted as canned debates and public spectacle is becoming unacceptable to an intelligent populace. New politics demands semantic understanding and identifying the chains of reasoning. These goals require building new tools and networks for the next generation of machine politics.

Back to Turing

One central issue is encryption, a topic with deep AI roots. Although Alan Turing is probably best known in computing circles for his Turing Machine and Test, he developed specialized electronic computation engines to decode German military code, which let the British withstand the Nazi air force. The current encryption controversy involves many players: the National Security Agency, FBI, NIST, telecommunications companies, software vendors such as General Magic, civil liberties advocates, the cypherpunks, and CPSR but no Nazis!

Meanwhile, information needed by the public for political analysis is increasingly available through the networks. A few examples: 1994 marked the first time the budget of the U.S. government is available in electronic format. You can now e-mail the president@whitehouse.gov. The Library of Congress is on-line. UseNet has a range of newsgroup topics, and users can post messages or merely lurk. All bulletin boards seem to be gravitating toward the Internet.

Networking also has proven to be an effective tool for grassroots organizing. For example, SeniorNet, the San Francisco, California-based network with health care as the prime concern, has more than 10,000 individuals. Several city governments for example, San Antonio are experimenting with the electronic town meeting discussion model, which was an idea originally proposed many years ago by Buckminster Fuller. Trade unions are going electronic, although there is no example yet of a union struggle being won or lost on the basis of electronic communication.

A key issue that remains is how to structure wide-scale electronic debate at various levels that is democratic, interactive, and inherently controversial. Can electronic discussion be organized and protected from dominance by lobbyists, special interest politicking, and the dirty politics of character assassination and mudslinging, while protecting the right of free speech?
One environment that begs for experimentation in political exchange and consensus is the World Wide Web, a distributed hypertext-based information system. With viewers like Netscape and Mosaic to unburden users with the technical details, users can focus on interacting with the information itself.

Omnicompetent Citizen

In 1925, Walter Lippman observed, "Although public business is my main interest, I cannot find time to do what is expected of me in the theory of democracy; that is, to know what is going on and to have an opinion worth expressing in every question which confronts a self-governing community. And I have not met anybody, from a President of the United States to a professor of political science, who came anywhere near to embodying the ideal of the sovereign and omnicompetent citizen."

Lippman's observation still rings true today. Does the public really want a daily digest of political information? An omnious trend toward political dysfunction is that the number who vote in national elections continues to slide below fifty percent of the eligible voting-age population. One possible reason for this trend is that many people believe that political representatives have little to offer in terms of solving the immediate daily concerns of employment, health care, education, housing, transportation, drugs, crime, social decay, injustice, and so on. Maybe, if the right tools were available, people would have a better chance to communicate with representatives, know and protect their own rights, engage in deliberation, test hypotheses, discover knowledge, discuss theory, and better understand world events.

Obviously, merit exists in the public becoming more politically astute and "awakening from the dormant state." Success may depend partially on whether participation can be achieved in such a way as to impinge minimally upon the matters of private life. This "awakening" is the challenge for a politics of knowledge. Advanced information systems at least may put the right tools on the table.

It will happen this way: Imagine a hot afternoon at a future Fourth of July picnic, when you are telling your friends about the details of a new Pelican Brief theory, debating the merits of admitting Cuba as the fifty-first state, or discussing the ramifications of Charles Barkley running for U.S. President . . . nah, please pass the potato salad!

Full Text COPYRIGHT Miller Freeman Publications 1994