[aiethics] recap of 1st ammendment thread from Alice users

[rdreyer@cfl.rr.com]

Jonathan Brody wrote:

> Perhaps my messages got lost, but a few weeks ago I raised the=20
issue of the
> <gossip> function exposing botmasters to liability for disseminating
> libelous speech. If someone told my bot "My ex-girlfriend, named=20
xxxx, who

I am hoping to bring this topic up with the Alice AI Foundation legal=20
cousel
within the next week or so.  Obviously there is a point where fun=20
software and
items which are very unique to create get lost on what the right=20
thing to do,
whats possible, and what construdes as disregard for users privacy. =20
Just
because there are no "laws" for bots @ the moment doesnt mean that=20
someone isn't
liable for them, even if the open source mantra of GPL claims "you=20
this software
as-is" with no warranties and no guarantees.  It obviously fits into=20
its own
class because Im sure that looking at what an open source library for=20
a part of
something and an open source interface for a user/machine interaction=20
are not
the same.

Not saying this now, since we all know outright that chatting on a=20
webpage
advertised as a bot is ok, but getting instant messages from bots and=20
email may
also constitute as "intelligent spam".

Imagine Alice as a telemarketer?  I think its hard for me as well=20
because as a
developer I know that there would be legal limitations in justifying=20
why a bot
would be good in the right hards and bad in the wrong hands.  Alice=20
could have
*easily* been able to wrap up audio files and be a Napster like=20
application but
is the headache really worth it?

It should also be viewed as "scientific work and research" first and=20
foremost I
think,  which really is a contradiction on my part because of the=20
work Im doing
to get an "Alice Service Provider" up and running.  The line is so=20
thin on where
it fits.  I think your case for <gossip> as a legal issue is full of=20
merit to
ponder about and worth asking what the legal ramifications are, if=20
any.

- Jon (not a lawyer)
_________________________________________________
Actually the worst that could happen is they'd ask you to remove the
specific line from your gossip file. I would imagine Alice's gossip=20
feature
to fall under the same laws that govern chat rooms and message=20
boards. If a
message board is deemed to be obscene you just have to remove the=20
material,
not close the service.

-Dekker

________________________________________
You think so? Read this=20
(http://www.salon.com/tech/feature/2001/07/13/museum_security_network/
index.html)=20
first.


-Jonathon
________________________________________
I very much hold to my prior statements. The case explained in the=20
Salon
article has a fundamental difference between itself and the Alice=20
gossip
feature and that is that the email was forwarded by a human being. In=20
the
case of alice you're talking about the exploitation of a software=20
system,
not the willful intent of a human being to post or forward unconfirmed
information or information of a personal nature.

"Should the landmark 1996 Telecommunications Act, which protects=20
Internet
Service Providers like AOL from libel in the case of third-party=20
postings,
also cover an e-mail newsletter that publishes unedited letters and=20
press
clippings? " - Jori Finkel , Salon.com

This is shaky ground for newsletters, because the issue at point is=20
that
this man, a human being with human thought processes, forwarded on=20
this
information to 1000's of individuals. Now from the stand point of=20
ALICE we
can see the chatbot falling easily under the blanket of an ISP's=20
service,
therefore it is protected. This is demonstrated by the fact that=20
ALICE is a
fully automated internet application and that the feature in question=20
has no
ability to discriminate or warn the system administrator of possible
slanderous remarks. The only way a human would ever even know such a
statement was logged is by performing a routine check of the gossip=20
file
(upon which time most material of this nature could and should be=20
removed)
or by a user complaining of the nature of the comment. The exact same=20
rules
that apply to those hosting a traditional chat forum.

Similarly, you wouldn't seriously expect to be held liable for=20
inputting

<pattern>*</pattern>
<template>go for it</template>

and having someone consult the advice of your chatbot on whether he=20
should
murder his family.

_______________________________________
Although it's true that alice is a software program and can not act
willfully and with malice, the Aiml author could potentially be held=20
libel
in the case that he/she programmed the bot to act in a malicious or=20
reckless
manor. It's just like a gun - "Guns don't go to prison, people do".

<pattern>*</pattern>
<template>go for it</template>
Putting this particular aiml on the internet could also be considered=20
to be
reckless. If some big corporation posted this aiml on the internet,=20
it could
be really embarrassing to find themselves in a lawsuit where the=20
defense is
claiming that the defendant murdered his family after being urged to=20
do so
by the corporations online bot. It sounds crazy, but this is just the=20
kind
of way-out thing a defense attorney is looking for. The press would=20
probably
love it, and the botmaster might be our of a job.

These are just my opinions. If you are concerned about this matter,=20
talk
with an attorney.

Conan




=20