Neurotech governance

From Wise Nano

Can better governance be achieved by requiring that people in government undergo some form of neuropsychological treatment?

Background for discussion

Assume for the moment that it will become possible for people to diagnose and change their mental processes and motivations. Labels like "sadist" and "control freak" and "altruist" and "nurturer" and "liar" will be understood in terms of their physical neural implementations. People will be able to change themselves to lie without guilt or stress--or to be nearly incapable of lying.

Key questions: can these capabilities be developed to an adequate degree pre MNT? Are they likely to be? How long will they take to develop with MNT? Can we stay alive for that long?

As this technology is developed, it will be used. Corporations would have to use it, at least at the executive level, or be outcompeted. So the question is not whether it should be used at all, but how it can be used for improvement. Do we want today's politicians in today's systems "improving" themselves to win elections and increase their personal power? People who are dishonest until caught (and even afterwards--who cares these days about broken political promises?), greedy, control freaks?

I'm not sure about this. The last 2600 years, and especially the last century, are full of technologies for mental self-modification which have almost consistantly failed to enter widespread use. This includes everything from good neumotic tricks to hypnotism to biofeedback to meditation to speed-reading to widespread math-literacy to tools for checking an argument's logical consistancy to universal languages (natural or artificial) to procedures for aggregating data, providing incentives, optimizing solutions, and imposing accountability. It includes meditative skills that can be learned in a few hours, nutritional and sanitary knowledge that can be summarized in a pamphlet, and thirteen year full time mandatory educational systems which fail to instill any of the above. Self-enhancement techniques tend not to be good memes. I don't know why this is, but it's obviously true. Corporate executives may make somewhat greater use of them than the general society, but they don't seem to be competent at recognizing and taking advantage of the best techniques available. I don't see any reason to expect radical sudden change in this respect due to MNT. Related question, are there ANY organizations currently devoted to testing, developing, and utilizing this sort of brainware? If not, why not? If so, why do they lack impact?

Or would we rather see a different set of improvements aimed at good governance? People who are guaranteed not to be motivated by sadism, personal power over others, or the desire for personal gain from being in politics (including alpha-male posturing and attention-seeking)?

I don't think that the citizens of a Democratic nation would be likely to prefer someone who set out to become radically different from them, even if the difference was in terms of increased altruism, rationality, and reliability. After all, they don't seem to want reliable rational people or altruists. Look who they vote for. I suspect that people have an evolved reluctance to trust those much smarter or more capable than themselves. A change of context won't overcome this reluctance to place their fates in the hands of a power than they fundamentally cannot understand.

This article was motivated by comments to a post on CRN's blog. "John B" asked how the "improvements" would be decided on. It's a fair question. Is there a mindset, or combination of mindsets, that promotes good governance?

Deciding the goals

Who decides the overall goals? If the goals simply evolve without direction, they're likely to be dangerously short-sighted. An organization that got better and better at focusing on a short-sighted goal could develop into quite a problem. (The same reasoning applies to Runaway AI.) So one of the goals, perhaps the main meta-goal, would have to be to limit the scope of the organization.

How could the mental recipe be chosen? (Again, we're assuming the technology is available to implement the recipe.) Start with a panel of experts: neuropsychologists, political scientists, sociologists, maybe ethicists--and a few random "hard scientists" to help the group avoid fuzzy thinking. And the first thing they'd all have to do is increase their honesty, decrease their ego, and increase their communication skills and intelligence. Then they'd get to work developing a mental recipe for good governance. (They might start by designing a better group for the job.) Once a government was in place, one of its tasks would be to continually improve itself. So the recipe would evolve over time.

The above panel was of course selected because it appears to a modern academically oriented and somewhat technocratic Western liberal to consist of the natural set of legitimate authority figures. People of other cultures might replace the ethicists with theologians, the sociologists with preachers or mullahs, and the political scientists with aristocrats or marxist or nationalist historians. Would this lead to a desirable outcome? Ego, honesty, and communication skills are not likely to be neurologically simple. They are also probably not easy to measure, especially ego. Different ways of boosting measurable intelligence may have different side effects, and some of these side effects may be what we mean by "intelligence".

Are there any goals that people can agree on? A major choice is between diversity and conformity. Lots of people have wanted to apply "central planning" to human happiness. But "central planning" appears to be a short-sighted goal of the kind that can lead to vicious cycles. The end result is Madeleine L'Engel's Camazotz: religious totalitarianism, in the worst sense of "religious": form that crushes spirit.* So it seems that a good government must act to preserve diversity. In fact, "preserve diversity" is quite broadly useful as a rule of thumb. Many crimes have the effect of hurting someone else's ability to self-express. Self-expression is good, until one person's self-expression damages another person; then, if civility can't settle the conflict, law must step in.

Many people do NOT want to preserve diversity. In fact, most of us are inconsistant in this regard. We want diversity AND the universiality of our conception of rights, justice, etc, but liberty is in some respects at odds with diversity, hence the anti-modern backlash against globalization, which increases liberty but destroys diversity. Is it Diverse for people to be raised as fundamentalists? Is it diverse for deaf parents to have deaf children? Is it diverse for there to continue to be mentally retarded people? To some degree we want to grant our liberty to others, yet by doing that we destroy what they are. I would suggest that preserving diversity is a goal with too great a scope to realistically expect to express with our current knowledge, culture, experience, and intelligence, without the goal becoming a runaway system. This proposal is essentially, as you have said, a proposal to build an AI out of people. We know a little more about how to do this than about how to build one out of transistors, but not nearly enough to propose safe open-ended goals.

• Form, even rigorous form, does not always crush spirit. Ritual, yoga, and so on can create space for spirit to flourish. But for this to work, it has to be voluntary, not forced: the practitioner must have responsibility for the practice. Some religious traditions hold that submission is desirable, but involuntary submission is an oxymoron.

Implementation

A major worry about any government is: What if it gets subverted? What if it is deliberately corrupted by people who are inimical to its purpose? Part of the mental recipe would include subversion resistance: no bribe-taking, no putting personal preference ahead of the public good, and so on. This would work best if everyone in a position of governance were required to take the treatment.

Wouldn't total transparency (all activities by all officials recorded and universally available) be a less radical but nearly as effective and far safer way to prevent subversion? It would tend to also be much more Democratic, though if the authorities chose to I suppose they could have transparent lives and simply ignore public preferences.

Unless things go disastrously wrong, a well-designed government would stay open to input--including input on its own design. This raises the possibility that it will encounter or develop a bad idea, put it into practice, and spiral out of control. Openness, good communication, self-examination, dedication to well-chosen ideals, and willingness to admit that a decision was wrong will help to reduce the impact of bad ideas.

I'm not sure that the last sentence follows. If a top level goal is implemented which is bad from the perspective of previous goals, that goal will still be right from its own perspective, and will tent to self perpetuate. At any rate, this discussion has become so abstract that is is essentially isomorphic to [AI Friendlyness theory http://www.singinst.org/CFAI/index.html].

Political scientists, psychologists, sociologists, ethicists, anyone--please criticize these ideas!