This idea is related to UserRatings.
Goals
An opinion system is only as good as the entries in it. It also needs tons of ratings. How do you encorage people to do that? Make endering the opinions interactive and fun -- like a game.
The opinion system allow the user to enter natural language statements (powered by cyc) to give their opinion to things. If the user's opinion was ambiguous, the system could then ask the user to clarify, and it could continue until the opinion is fully understood (thus reducing the statistical noise as much as possible).
Example:
What would you like to give an opinion about? > my sister You want to give an opinion about a person, your sister? > yes Okay. What is her name? > andrea So, her name is Andrea <User's Last Name>? > yep! Okay. And what is your opinion about her? > she smells funny and never gets off the phone Please verify the following facts: Andrea <User's Last Name> ------------------------------------- 1) Has an unpleasant odour 2) Uses the phone excessively Is this correct? > yep! Your information has been recorded. You are one of 23984723987432 people who have commented about Andrea <User's Last Name>. The following statistics have been recorded: Andrea <User's Last Name> ------------------------------------- - Totally hot (global opinion: 54%, personal relevance: 0.01%) - Has an unpleasant odour (global opinion: 0.01%, personal relevance: 100%) - Uses the phone excessively (global opinion: 12%, personal relevance 100%)
- Opinions should be personally relevant (as shown above) by clustering people.
Issues
Spam
Companies are going to try to flood this with positive comments about their product. At first, it'll probably be manual. Then they'll figure out how to automate it, which will force us to filter out robots. Eventually it'll just be a squallid pit of crapulence like every other public opinion system.
Partial Solution: Use social networks (facebook, for example) to filter data to known humans. Blacklist "bad" people. People slowly accumulate cred over time for reviewing products honestly.
Collectivism
aka. groupthink or mob mentality
Jaron Lanier summarizes it quite well:
The collective is more likely to be smart when it isn't defining its own questions, when the goodness of an answer can be evaluated by a simple result (such as a single numeric value,) and when the information system which informs the collective is filtered by a quality control mechanism that relies on individuals to a high degree. Under those circumstances, a collective can be smarter than a person. Break any one of those conditions and the collective becomes unreliable or worse.
Meanwhile, an individual best achieves optimal stupidity on those rare occasions when one is both given substantial powers and insulated from the results of his or her actions.
History has shown us again and again that a hive mind is a cruel idiot when it runs on autopilot. Nasty hive mind outbursts have been flavored Maoist, Fascist, and religious, and these are only a small sampling. I don't see why there couldn't be future social disasters that appear suddenly under the cover of technological utopianism. If wikis are to gain any more influence they ought to be improved by mechanisms like the ones that have worked tolerably well in the pre-Internet world.
The hive mind should be thought of as a tool. Empowering the collective does not empower individuals — just the reverse is true. There can be useful feedback loops set up between individuals and the hive mind, but the hive mind is too chaotic to be fed back into itself.
Jaron's essays on the topic:
Edge.org: Beware the Online Collective (by Jaron Lanier)
Edge.org: Digital Maoism: The Hazards of the New Online Collectivism (by Jaron Lanier)
Opinion "weight" is unknown
A paragraph written by Einstein is going to be intellectually "denser" than a paragraph written by a 14 year old blogger. They could both contain the same number of words, use the same grammar and syntax, and even similar vocabulary, yet the concentration of thought and information in Einstein's paragraph will be huge by comparison, since Einstein would spend hours pondering ideas until he had looked at them from every possible angle.
An opinion (or article) written by a person who's had years of experience (and thought) within a domain is of much greater "weight" than that of a 14 year old who's talking out of his ass.
What's required is more information about a person's background and achievements. It's not perfect, but it's a rough guide.
Partial Solution: Social networks (again) allow you to learn about people and verify reliability.