As advertisers look to better target ads they might actually be getting closer to creeping people out. Might the same be true for the dream of AI powered intelligent agents in virtual worlds?
Ever since GeoCities sold for billions and thereafter faded away, it’s been known that it’s difficult to profit handsomely off the back of user generated content.
Yahoo Answers doesn’t bring in the premium advertising rates that other properties command, so it isn’t crazy that they’d try to sell it if the price was right.
User generated content equals questionable content and advertisers don’t want their ads showing up next to it. User generated content also doesn’t seem to click as well – the theory goes that people are only there for the content and not in the mindset to click ads. Though advertisers always hold out hope that better targeting will change all that. Next time you visit Facebook it will know that it’s lunch time and that you tend to contact certain friends around this time – perhaps you’re asking people about lunch? Wouldn’t it be nice to see an add for Taco Bell or a coupon? 
The scenarios go even further, some of which Seth Godin covers when he likens this dream to the uncanny valley – where as robots and 3D graphics look and act almost lifelike that the creepier they get. I think the analogy is fitting. Even if the use case would seem helpful, at times it would just be creepy.
The same is true for those dreaming of making AI to power intelligent virtual agents. Some imagine a virtual butler in virtual worlds. This virtual butler would assist you with scouring the web for interesting news just like your RSS reader or Google Alerts, however they would do even more. They would have learned that on Fridays you care more about movie stories and less about the stock market. They would tailor the collected stories to your liking. However the better the engineers make the AI, the closer they may be getting to the uncanny valley and creeping us out.
Related posts:
A Senior Product Manager at 