[ExI] free-will, determinism, crime and punishment (& CS techniques - Security and/or Privacy )

Eugen Leitl eugen at leitl.org
Mon Sep 3 04:57:39 UTC 2007

On Mon, Sep 03, 2007 at 02:15:48AM +0300, Khaled Aly wrote:

> I've been following this since it started. I feel that a murderer should be 

That thread has been killed, actually.

> punished to death in the easiest possible way (e.g. lethal injection), once 

You don't seem to know much about execution by lethal injection.
Especially, since many are bungled on purpose. Say, have you ever been
in jail? 

In terms of monetary costs, executing people is more expensive than
locking them up for life.

> it is proven for very sure that he/she has done so deliberately and under no 

You don't seem to know much about 'proving' anything outside of the realm
of formal system. How do you estimate your error range in knowledge?

> external effects; for two reasons: 1) It is a relief to the victim's family 
> unless they choose otherwise; and 2) If I were to choose, being the 

Do you know how clan wars start?

> murderer, I'd rather go now/then than spend 20 years in jail. About therapy, 

Since you think that way, everyone else must also think that way. Obviously.

> I think it may work in certain sick cases (there are sick souls that could 
> be treated why not, and there are evil souls that will not change, and there 
> are those whose life circumstances made them go wrong way and they need 
> social rehabilitation - every case is different).

Interesting theories you got going there.
> This is my entry to how effective is the overall current justice system. And 
> the ever unresolved question about the tradeoffs between personal privacy 
> and community security. Do any computer or IT people see a role of 
> algorithms helping the justice system to decide (for a start; since 

The legal code is already an algorithm by which the society operates
(and it is really code in literal sense of the word). Fortunately, blind
Justitia is executed by agents of flesh and blood, which have common
sense (the law is an ass).

> real-life AI is a bit far ahead). What I mean for example, consider 
> obtaining a search warrant. How difficult is it to write some code that 
> could 'assist' the decision making. First, it will perform preliminary 

If people need an expert system to decide such basics you should fire them.

> information analysis. Second, if it's open implementation, it will be 
> possible to a large technical population to review it and ensure it works as 
> intended/declared and for the benefit of justice. And third, within a 
> digital world, it can ensure that a court order of privacy invasion for a 
> suspect individual will actually expire; provided it began, through the use 

People make mistakes.

> of digital certificates to be provided and revoked in time by court to law 
> enforcement. The last of course requires that default electronic 
> communication be secured, and be broken only using a court digital 

Good idea, in theory, in practice cryptography doesn't work.

> certificate. This can be as frequent as the low issuing entities would 
> decide according to existing situation (exceptional, less exceptional, 
> regular). An innocent whose privacy was broken deserves to be advised about 
> it at some point.
> Beyond this, any computer literate person would confirm that open source 
> code is most reliable because it had been reviewed by the expert public. 

Any code is unreliable, read Bugtraq.

> Same applies to an open source algorithm that is designed to support the 
> justice system deciding what to do with a suspect, a person in trial, or an 
> indicted person and let that be an advisory input for trial.
> Unlike few people may think, computers and human minds don't work the same 


> way. Computers are structured crunching speed machines and minds are pattern 
> learning neural machines. They are complementary (until neural network 
> computing make it to market and yet it will be v. hard to mimic the brain). 
> Why does law enforcement, or as a whole the justice system use computers 
> only for data recording purposes. What about the huge analytic power that 

Do you know many LEOs or judges? Do you know much about system security?

> can be made presentable to humans to evaluate and judge. I don't think any 
> final sentence should be produced by a machine with the current state of the 

Thank you for that.

> art, or may be never. But speaking preliminary and analytic support, 
> incorporating digital technology can potentially resolve many issues and 
> cases where the society disagrees about how just the outcome was.

I could see whether computer assist would be good in forensics, but in
decision-making, that's ridiculous. By the time it will be useful, there
be computer crime, as in: criminal computers.
> Greatest software technologies came out of open source because it is 
> auditable. And that's much like digital democracy. Transparency does not 

It is so difficult to build robust electronic voting systems and people
are so ignorant that currently all such attempts need to be banned.

> compromise rule of law -- it rather enhances it. Sadly, digital technology 
> is being consumed for many irrelevant but sellable applications before it is 
> being considered to support a sound social infrastructure. And what's more 
> relevant than justice...
> ka, phd

You have no idea how funny you are.

Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE

More information about the extropy-chat mailing list