You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Would it be possible to have an AI set up to only help advance a government and the will of the people? Not to replace all politicians or even have a vote, but to clarify the viability of projects and legislative actions. Looking at the quality of politicians, this might be a viable option. All code would have to open and accessible for public viewing. It couldn't be worse than some of the politicians we have and could be run open book for all to see how it functions and reaches decisions. Any thoughts and views on this? The quality and integrity of politicians isn't to high at this point.
In the book series "Revelation Space" by Alastair Reynolds*, AI is used to enable a "live democracy" where everyone in the population participates, and on a regular basis. The votes are proposed via neural link, and voters are required to weigh in. The voters can also achieve a weight of up to 4 votes if their previous votes were determined to be the correct ones. (Yes, very sinister sounding, thought the book talks about long term votes such as how to deal with threats to society, alien species engagements, etc).
Not saying we should do this btw just sharing as an example.
I think AI *should* be used for transparency to voters of politiicans actions and educating politicians on the impacts of their choices. Ex: reporting on the possible conflicts of interest should they vote 'yay' or 'nay' on a bill, or calling out potential unethical actions (i.e. see the Nancy Pelosi stock tracker - is she insider trading or not? ).
*This topic is covered in a bit of detail in the "Prefect Dreyfus" series - a later 3 part book series set in Revelation Space.
I am confused that I hear reports of AI being so accurate and powerful. In the past month AI has told me.
1) Mallory Swanson (soccer player) has a child already (reference them announcing a dog a few years ago as "their baby" on social media, with a picture of a dog). She is expecting her first this year
2) July 26th 2025. is a sunday (I was googling a wedding, AI cannot seem to read a calendar). it is a saturday FYI
3) Three different solutions for a problem I was having with my accounting software. None of the prompts or buttons it recommended actually exist. It did say "Im sorry, your are correct, that button doesn't exist" after each time.
4) It told me I can walk downtown and stroll around from a house 10 miles from downtown (Real estate ad). It was referencing an advertisement for the city and putting it in my ad.
These are all facts.... I won't even get into when it comes to solutions for crime, economy or "Is it good to change tariffs daily based on your mood?"
As long as everyone realizing anyone can make a CGI animation to show anything, I am not worried.
It does help make it so that googling something takes 10 seconds instead of 1 minute to get an overview. I would not Reference any summary from AI/Chat GPT
I have started to see AI generated baby themed interviews of politicians and football pundits which are quite entertaining .... if 100's of data centers are being built for the likes of this and if we are fast tracking nuclear power plants to feed these energy hungry data centers so we can see more polished AI generated baby themed interviews ... well then its on us
I was going to a major sports final last weekend, I asked ChatGPT how many times one of the teams had one the title and ChatGPT told me how many times and included that the team I mentioned had also won it in 2025, a full 24hrs before the final had taken place ... ... unfortunatley they lost the title!!!!
). 
... unfortunatley they lost the title!!!!