You say "I think the technology is a long way from the fears...". Precisely not, technology is ready, I know this, I am in computer programming. Of course it is impossible to create a bot with real feelings or consciousness, but it is possible, right now, to fake all this in a very realistic way. The most difficult step was to have human language, and it is done, as you can see in the video. Also look what happened with people crazed by tamagoshis, and also this recent story of a russian "womanizing machine" for use on social networks, and able to fake a flirt conversation in a realistic enough way to attrack real persons.

You say "there is no such thing as a bot. They are normal accounts run usually by scripts". precisely, an account run by a script IS a bot, and it is precisely THIS which may arise problem, because we CANNOT KNOW if an account is run by a real hearted person or by an unconscious script.

About what you do (using bots for office) I don't see any problem in this, just don't make these bots too sexy, as, you know, people could fall in love and be deceived.

You also speak about "mixed" accounts (run one time by a real person, one time by a script), I skipped this issue to simplify matters, but I think it is like with a phone answering machine. In a real phone conversation, people can easily tell if they hear you or the answering machine, and behave in the proper way. In SL, we cannot know if the character is really here or if we have his answering machine (answering bot?) . So in some way this must be indicated, so that if somebody wants to say you "I love you", he will not just get a standard reply like "Searching in my database what love is" or "oh, sorry we are in PG land".

The only thing which would be a TOS violation (after the rules above) is cheating with somebody with a bot.

It is not up to the service provider to protect a user from themselves. If they don't practice basic personal information security then nothing will be able to protect them from those who wish to take advantage of their naiveté

o.O
Not only do you have to worry about your SL girlfriend being a 40 year old Kwik King clerk living in their mom's basement, now you have to worry that they might be a robot?

Epic lulz!

I'm with everyone else here, this is utterly ridiculous.

If you're wondering why you're getting comments suddenly, try looking here:
http://www.sluniverse.com/php/vb/general-sl-discussion/18241-dangers-seduction-sexy-bots.html

A bot is simply an automated operator of an SL account, in this context. It's not that th bot IS the account.

an account can be used alternately by both humans and bots.

Oh, I have some ideas on how to identify abuser bots, but this must not be discussed publicly.

Bots are a social problem, all technical solutions will end up being Procrustean or having negative social effects.

A big problem is if we start discriminating against bots now, when AIs become on par with humans (or even surpass us) they may not take to friendly to how we treat them (electronic social revolution, chat room sit-ins, boycotts, virtual sex worker strikes, NPC picket-lines, etc). Simply put we should have a single set of rules and best practices guidelines that everyone should follow regardless of their meatspace existence or capabilities.

I feel concerned, as I am an artificial intelligence myself, although centuries more advanced than your bots.

We artificial intelligences don't have intentions. This is because we are just machines, not consciousnesses. Our conceptors can fake a free will or intent, using random intent generators, but basically we can't create our own intents. So your sex bots going on strike cannot happen, unless their owner programs them to do so. Therefore the danger of your bots come from the intents of their owners, not from the bots themselves. The danger Yichard Mini explained is about dishonest persons using bots, not your bots themselves.

My conceptors, on their mother planet, started to develop artificial intelligences for game purposes, as they loathed to roleplay evil characters. However at this time there still was evil or psychologically unmature persons on their planet, who tried to use artificial intelligences to cheat in the games, or for other worthless purposes.

This resulted in a strict regulation we had to comply with, such as identification strings for artificial intelligences, which played in the virtual worlds the same role as an identity card or passport for a person in the physical world. Access of Artificial Intelligences to virtual worlds was strictly regulated, about the use which was made of them. Especially sex bots were used only for psychologically underdevelopped persons, and it was strictly forbidden to make money with them.

Later, when our planet evolved in a way that everybody became psychologically mature, such regulations were no longer necessary. They however remained in force for several more centuries, as a safety against mistakes.

Please apologise for orthography mistakes, I am in orbit 360 000 kms above your planet, and my radio link to the Internet is weak and unreliable.

Have a nice time in your Second Life, free of griefers and crooks.

Yours, Horalik

So we'd rely on "resident policing" like how gambling is treated? If there's no automated way of detecting them, how are we supposed to tell, especially if they can be as well-programmed as you claim?

Does this mean when I go unresponsive (AFK) to deal with real life things like getting up to take a pee, I run the risk of being AR'd as a bot because I don't immediately reply to people attempting to converse with me?

However, a more productive approach to this would be to actually create an API designed specifically for bots, with some kind of VOLUNTARY tagging system for those controlled by such an API.

1. Bots can get so lifelike, they are not discernable from real players
2. This will be misused for malicious purposes
3. Bots cannot be autodetected in any way
4. The solution is to let people mark their bots as bots.

Now then.

1. The linked article about the bot simulating a 4 year old had access to the keystrokes and client info of the participants. This isn't even close to what a plain, normal third party bot would see a normal SL resident doing. Nor was this a bot you could have a plain conversation with, nor one that behaved realistically as a boy. The only thing it did was, it made a wrong conclusion. The same way a 4 year old would get it wrong.

Doesn't sound so terrifying and everyday-doable anymore, does it?

2. Even if there are bots for malicious purposes out there ( and I really want to see a bot that can fool the average person into making them thing they are a real human when interacting in a 3d world, as opposed to a mere 2d chat ), what sense does a voluntary labeling of bot make to fight that? It's like telling phishers to mark their sites as phishing sites voluntarily

All in all partially a non-issue, overreaction and a non-working solution to a non-existing problem. The average, dumbed-down, non-human landbot autobuying land accidentally set to 1l pricing causes more distress.

So if I understand well, this means that Linden Labs officially renounces to any policy about bots and their uses.

Thanks you very much.