Scene 11: Barry and KTS

In which Barry goes to KTS to get smarter about an unfamiliar topic, and gets impish.

Users of personal learning assistant systems will not be manipulated when they do not wish to be.

Barry had a lot of respect and gratitude towards Malik. Malik's wife had helped Barry's wife Sara through a long illness. And though (it seemed to Barry) Malik was a lot smarter than him, Malik was always very attentive and respectful towards Barry.

After a Sunday service, Malik invited Barry and Sara to stop by their house for coffee, and the two men found their way into Malik's back yard. The conversation turned to politics and to recent news about candidates in the upcoming presidential election. Malik planned to vote for Pieters, the Globalist, and had been making points that riled Barry inside. The conversation became a bit awkward, though still friendly.

"I know that countries gotta play by the rules, and that we're interconnected and all that," said Barry, "but haven't they proved that… that the lack of accountability, you know, of the UN, causes individual countries to get screwed."

"Sure," replied Malik. "National self-determination, very important. There needs to be power at every different level, it's hard to get it exactly right."

Barry grunted. "Sure. I guess—I guess I think most of the power should be national and state. I don't like decisions being made for us way up there. Decisions we don't even understand or know about!"

Malik nodded in agreement. "Yes, we have to ask questions, we can't let them get out of control."

Barry smirked. "But you don't think they do get out of control."

Malik laughed. "You have to judge each issue, I guess. Like, I think some of the restoration, the environment restoration is too ambitious. Other things, like in Africa, I don't think they're doing enough."

Barry recognized that Africa was a big empty area in his mind.

When he got back home later, he unceremoniously said to Know the System, "Explain Africa to me."

KTS considered Barry's recently-renewed interest in things political, but legitimately needed a bit more direction than that. And so it queried:

Is your focus:

a) The candidates' policies towards Africa?
b) African challenges?
c) African geography?
d) Something else?


Barry picked "b", and a ten-minute video launched (titled, somewhat amusingly, "Executive Briefing".) It presented some demographic information—population density, income, education—compared to other global regions. Barry was impatient, but let it play. Next, it provided a condensed history, pre-colonialism to the present. It then started describing some of the countries and their difficulties, using the phrase "weak state" several times. Finally Barry said "stop," and, automatically, a handful of frequently-asked questions appeared.

He stroked his chin for a moment, then selected one of the questions:"What are the elements of a well-functioning national political system?" An animation began.

Scott in den
"A modern political system consists of three sets of institutions: one, a modern, impersonal state; two, the rule of law; and three, mechanisms of democratic accountability.

"A 'state' is defined as a legitimate monopoly of force over a defined territory. Pre-modern states were 'patrimonial,' which means that the rulers considered the state as their private property which they could use for their own purposes. A modern state, in contrast, is impersonal, and corruption occurs when rulers take public resources for their own private benefit."

"Stop." This was good stuff, if a bit complicated. Already, though, he thought he could see some of the reasons why the Globalists were wrong and the Libertarians were right. He stumbled forward. "Do Globalists… cause corruption… by, by… ignoring democratic accountability?"

Is your aim:

a) to win an argument with someone
b) to get the quick Libertarian explanation
c) you'd like to devote some time to learn about this


Normally these choices were the ones Barry liked, and usually one of the three matched his relevant desire. This time, though, he felt oddly annoyed, and wanted to knock KTS around a bit. He asked, "Why do you think I would want the quick Libertarian explanation?"

KTS's deeper algorithms sprang into gear. It considered and then quickly rejected a potential reply that Barry had previously affirmed that he always wanted that choice included there. Its general aim in situations where Barry seemed argumentative was to be honest, helpful, and avoid any possible insinuation that Barry wasn't being fair or straightforward. Barry knew this, of course, and—only occasionally—obtained some entertainment by tormenting his silicon assistant.

After many processing cycles (but only a couple seconds), KTS replied,

A couple possibilities:

  • You are trying to formulate your own policy stance regarding Africa, and you simply want to determine where Libertarian thought intersects most directly with Africa policy. Or,
  • You suspect you have discovered a solid, damning criticism of Globalist ideology and you want to evaluate more closely.

Is it related to either of these?

Something that we don't yet see much of, is the computer having a "theory" of what a person wants, and a "strategy" for getting them quickly to it. Today’s search capability returns to you a prioritized list of items that, based on the search words you entered, might be what you want. But what system lets you say "No, I'm not interested in X, I'm trying to achieve goal G" ? What system knows you well enough to predict—in general terms—what you are probably trying to do, and conversationally zeroes in on it together with you?

We won't see this kind of thing in the near term, except in very limited domains, e.g. a software app that helps you use that app. But it will be the general evolutionary direction, and it will be wonderful.

It is not, unfortunately, a realm where a new breakthrough technology is likely to yield free lunches. A lot of new kinds of metadata will have to be entered by developers on the content side; indeed the notion of "content development" itself will need to become more sophisticated. And the sheer amount of content needs to be greater; there needs to be content available for every relevant circumstance. Further, users will need to provide more information about themselves, and new kinds of user modeling will need to be designed by computer scientists.

Nevertheless, techniques are certain to accumulate, new tools and ecosystems will emerge, and helpful new technical standards will be formulated.

As depicted in this scene, Know The System could be very helpful to each of us in reasoning about our own political beliefs. Certainly it won't eliminate all "echo chambers" in which we prefer content that is biased towards our own biases (or, for that matter, biased towards news that is sensationalist, salacious, or celebrity brawling.) But if Know The System were "smart" enough—both about the philosophical debates and about the individual person—I think most of us would be enticed to scrutinize our assumptions more often because it is interesting, satisfying, and entertaining.

Notice also the “standard” three choices, paraphrased: (1) I need to win an argument; (2) I want my “team’s” position; and (3) I want to learn. These remind the user that there is a difference between protecting your beliefs and seeking truth, which may encourage a more open-minded attitude. From society’s standpoint, that would tend to limit or at least reduce the problem of political polarization.

The net societal result would be voters with more nuanced views about political ideology. I believe the category of "moderates" (as defined using today's standards) would expand to include the vast majority of the electorate.

By the way… the thought of an AI-based assistant conjures for some the specter of mind control. For example, Yuval Noah Harari predicts that future AI will be able to push people’s buttons so perfectly that free will may effectively disappear. That to me is laughable – absent some kind of terrorist government, people cannot be compelled to believe anything that they don’t want to. If anything, human “motivated reasoning” means that beliefs reflexively harden when contradictory information is presented.


cast of characters