Re: GPL on AI generated code

[Alexandre Oliva]

Hello, Akira,

On Aug 14, 2023, Akira Urushibata <afu@wta.att.ne.jp> wrote:

> The machines may be rigged to give a certain candidate advantages.

This is unfortunately true even if they run freedom-respecting
(transparent and obedient) software.  There's no way for voters to tell,
at the time of voting, whether the machines are running binaries that
correspond to published sources, even if they're published.

The solution for this is not voting with machines running free software,
it's robust voting protocols with human-understandable and
human-verifiable audit trails.  Computers do not provide that, they're
rather one of the several untrustworthy components, and one that cannot
be verified.

Even freedom-respecting software and hardware wouldn't cut it: they'd be
obedient to the *user*, by definition, and the *user* is not the
*voter*, but the electoral *committee*.


OT material follows.

Robust voting protocols enable voters to trust the process and the
results even without trusting the committee, but relying exclusively on
computers for any part of the process requires placing infinite trust on
them and on those who make them.  For trustworthy elections, we need
processes that produce trustworthy outcomes even if we distrust the
machines (if we use machines at all), and even if those who made or ran
the machines or any other part of the process were violent terrorists
who would pursue, torture and exterminate whoever they could determine
to have voted for their opponents.

> Donald Trump claimed in 2020 that voting machines were rigged.

So did Bolsonaro in Brazil in 2018, and again 2022, to the point that
the minions who invaded the Brazilian congress, supreme court and
presidential office on Jan 8th this year held signs demanding the source
code for the voting machines, in Portuguese and English!

But joining either side of this dispute is a bit of a distraction, and a
very dangerous one.

Claims that the machines are vulnerable are not disputed by demanding
evidence of fraud.  It is possible for the machines and for the voting
protocols to be vulnerable even in the absence of fraud, or even
suspicion thereof, let alone evidence.

Furthermore, evidence of fraud may be impossible to get when the voting
machines are designed to make fraud possible, massive and undetectable.

Absence of evidence is not evidence of absence, and absence of evidence
of one occurrence is not evidence of impossibility of any.


I hold a conspiracy theory :-) that Trump and Bolsonaro are both
denouncing the vulnerabilities of the voting machines under the
influence of groups who wish to ultimately *discredit* the notion that
they are vulnerable, so that, when these groups *do* manage to defraud
elections (in their favor, presumably), people won't give credence to
that possibility, and people and organizations who would be expected to
denounce their fraud would have already discredited themselves by
joining the groundless opposition to their groundless claims.

Would those who believed the potential denouncers' previous claims of
invulnerability of the process start believing opposing claims so much
like those they disputed before?

Would voting scientists who know better about the vulnerabilities of the
present voting machines have got used to disregarding the potential
denouncers' claims as unfounded?

It's quite machiavellian, if you ask me.

-- 
Alexandre Oliva, happy hacker                    https://FSFLA.org/blogs/lxo/
   Free Software Activist                           GNU Toolchain Engineer
Disinformation flourishes because many people care deeply about injustice but
very few check the facts.  Think Assange & Stallman.  The empires strike back