The community's fourth estate...Computers write stories about bills and laws that regular people can understand, using machine learning techniques.
RJacobsen is a professional software engineer working in Portland Oregon.
There are fewer and fewer journalists. American democracy is suffering from not having a functioning fourth estate. We need an open source project to teach computers to keep an eye on what lawmakers are doing and tell the community what we need to know to vote in our democracy.
The first objective is to scrape the web for bills from the Oregon Legislature and categorize them according to the categories listed on wikipedia's areas of law page https://en.wikipedia.org/wiki/List_of_areas_of_law.
Provide a subscription service so lawyers and other interested people can subscribe to and read the text of a law when it is posted on the web.
Expand to other states, local governments, and the federal government.
Have the computer write stories about each new bill. These stories should include a summary of what the bill says, who benefits and who loses, if there are any entities specifically named in the text of the bill. This task is all about the people and organizations affected by the bill.
Serve stories to subscribers who are interested in that kind of bill or area of law.
Track trends in bills becoming laws. Track trends in bills moving from state to state or municipality to municipality. Build on the capabilities which came before.
Answer questions about laws. Continue to build on what came before.
Inform the public when a bill is being considered by their representatives. Especially communities who lack lobbyists and may not even know when they are effected by a bill.
The project requires more than one person to work on it. (Need to lower the bus factor, that is, if one person is hit by a bus others should be able to carry on.)
This program must be unbiased or it will fail in its mission to bring real information to the communities it aims to serve. It must not become a propaganda machine.
Finding other professional software engineers who understand machine learning and are practiced in making it a reality might be difficult.
Quality must be high. In order to be useful the program must provide interesting reading content or subscribers will get bored and stop subscribing.
This will only work with government entities (federal, states and municipalities) which put their bills on the web.
First version is being written in Python on Windows 10 using Visual Studio environment.