http://news.rpi.edu/update.do?artcenterkey=2410 this link and video shows an experiment recently made in Second Life by the "Rensselaer Polytechnic Institute" where we see a video of a child looking av animated by a computer program, and speaking and moving like a human-controlled av would do. Predictably, we shall soon have bots with intents, memories, and most of the logical features of the human mind, even choices, opinions and prejudices, so that it will soon be difficult to tell a bot from a real person. I think that this kind of researches are in the same time enthralling and frightening. In some cases, our experience in Second Life is a play, or technical (learning, etc) so that a bot would not harm. But in many occasions, places or groups, our experience is not based on logic and situations, but on emotions, friendship, subtle mind matters which cannot be reduced to logic. More, this experience takes its value from the very fact that it is shared by another human consciousness, and becomes senseless, see weird, if we are just discussing with an unconscious computer. I shiver with the idea of bots faking friendship or love, not to speak of ugly things such as bot prostitutes... Such bots would raise many serious and specific issues, such as making financial schemes, spam or griefing, or cheating in roleplay, by being 24/24h on line, or a single computer controlling several "independent" bots. Making automated bot recognition systems will not work (it is already difficult on the 2D Internet). Many people don't really master logic, and thus would miss the tests! For this reason, I think we should take the following measures: 1) a bot should be clearly identified as such, at the very first level of its account 2) any land owner, group owner, etc... should be able to forbid all bots. 3) any player should be alway able to know he is interacting with a bot. (today animations chat appear in green, but a bot accound appears as a real person) 4) rule 3 could be disabled by a land owner for specific purposes. (roleplay...) in a general way, of for some specific bots he controls. 5) the region statute bar (top of window) should indicate the current bot authorization (none, identified, invisible...) 6) when creating an account, declaring a bot as a human should be made a serious TOS violation, similar to disclosing privacy or cheating with RL identity