Metaverse Bots: How To Spot Them, What To Do With Them?
Do the Lindens plan to address the impact of bots on Traffic stats? After three unsuccessful attempts to get a reply (any reply) on that question from their PR representative, that remains unknown. Meantime, I discovered this Jira feature request vote from Haravikk Mistral, suggesting that Traffic be eliminated as a metric altogether, largely in protest against bot farms. Go here to review and vote on it, if that's your inclination.
A larger question remains, regardless of what happens to Traffic-- how exactly do you recognize a bot? In the novel that inspired Blade Runner, Philip K. Dick had a fanciful machine that did the job, called a Voight-Kampff. In Second Life, that's quickly becoming a practical necessity. But what's the best avatar-based Voigt-Kampff test?
When I wrote about the issue a few weeks ago, I happened on two sites that seemed to have bot farms. My unsophisticated Voight-Kampff was to go right up to apparent bots and say, "Are you a bot?" At one nightclub, I didn't get a single reply, leading me to conclude, "bot". But as I found out at the other site, that's not a reliable method. I eventually did get a reply from Residents who were just away from keyboard-- even though it sometimes took a few hours to respond. (Or in the case of some, 12+ hours.)
And in any event, what happens when the bot farmers program their bots to have minimal AI and conversational abilities, a technology which already exists? I can see the fun in not knowing if the avatar you're dancing with has a human being controlling her. But at some point, isn't there an ethical obligation for bot owners to clearly designate them as such?