An attempt to use AI to ingest and organize diversity of thought.
We're making some bold statements that will drive the design and development of this project:
- Diversity of thought is fundamental to decision making
- Truth does not exist, or at best, is not important
- Facts do not exist, or at best, are myopically applied if they do
- Groups are wiser than individuals
- Decentralized control critical to removing bias
Breadth of understanding and diversity of thinking is managed by your piers, your thought leaders, or your media
Social media is flawed Creates echo chamber of similar thinking people with similar content bias and similar commentary
Internet is flawed Hyperlinks only link to content that supports a presented point of view
Media is flawed The main outlets have consumer bias, small outlets have can have integrity, but trust concerns, small outlets can be heavily biased and factually far from common belief
The truth that drives decisions is extremely malleable and therefore easily manipulated
Perception of truth depends on a number of factors
- Real experience of subject
- Breadth of understanding
- Depth and concept linking
Facts suffer from the same problem as statistics. Very few things in science are actual laws. Facts often represent faith and trust in institutional knowledge rather than actual hard truth.
Presented with diversity of thinking
- Individuals can make bad decisions
- Groups tend to make good decisions on aggregate
- Can only overwhelm knowledge by generating a huge amount of content
- Assign ranking based on related idea, not actual idea (i.e. you find the related idea to be “helpful”, not biasing your view on the active idea)
Use AI to analyze content from a broad range of sites:
- News: mainstream to independent, across the political spectrum
- Scientific sites: government committees to highly biased corporately sponsored
- Government agencies: particularly offices of statistics
In all cases – diversity will come from a spectrum of content sources
- Political: from far right to far left
- Influence: from highly influenced to independent
- Geographical: global content from unstable and dictatorship to free societies
Use AI to attribute
- Sentiment: based on categorization, is the view in support or against
- Bias: rating of content bias based on known journalistic and linguistic standards – not bias based on partisanship
- Frequency: count the occurance of the categorization
- Following: based on user group and sentiment of that user group to the content