Posted on 10/07/2025 9:27:00 AM PDT by Twotone
JIM BANKS????? ONLY JIM BANKS??? IT SHOULD BE ALL REPUBLICANS!!!
I’m sure they were honest errors by the bureaucracy. Pffffft
The Census Bureau is part of the “Executive Branch” of the United States government.
I’m trying to remember who is in charge of the “Executive Branch” ... just can’t quite think of his name ...
“Errors” my butt. They knew exactly what they were doing.
This shows how corrupt our government has become and how feckless Republicans are. For the Census bureau to say there were errors and we know what they were but we can’t correct them is total BS and should not have been allowed to stand. The fact that Republicans allowed it to stand shows just how weak and corrupt they are too. If it had been Democrats who lost six seats, it would have been corrected in a heat beat.
THIS!!! EXACTLY!
Where are the Republicans?????????????????????????
Talk about a golden opportunity for corruption and Democrat cheating.
Privacy legislation stands in the way of this, but I asked ChatGpt a unique question:
With the current AI and LLM couldn’t we meet the law requirements and actually get some real transparency on the data?
Answer:
🧠 1. Technically, yes — modern AI can now do privacy-preserving tabulation far better than differential privacy alone
Differential privacy (the Census Bureau’s chosen method) was designed in the 2010s, before today’s AI tools existed.
It’s a blunt instrument: it adds random “noise” to protect privacy but often distorts the data in small communities.
Since then, privacy-preserving machine learning (PPML) has come a long way.
AI could use techniques such as:
Federated Learning — the model learns patterns across decentralized data without ever pulling individual records into one location.
Homomorphic Encryption — allows calculations on encrypted data without ever decrypting it. You can tabulate results without exposing personal info.
Synthetic Data Generation — AI can train on the original census and create a “look-alike” dataset that statistically mirrors reality but contains no real identities.
Differentially Private Neural Nets — instead of random noise, the model’s training process itself limits what can be “remembered” about individuals.
Put simply: modern AI could produce block-level accuracy while still protecting identity, because the privacy guardrails can be built into the learning algorithm instead of corrupting the output.
⚖️ 2. The bigger barrier: Title 13 and federal bureaucracy
Even with perfect AI, the Census Bureau is bound by Title 13 of the U.S. Code, which states that:
“No publication whereby the data furnished by any particular establishment or individual can be identified shall be made.”
That language was written in 1954, long before encrypted computation or synthetic AI models existed.
So, legally, the Bureau can’t simply say, “We used AI to guarantee privacy, so we’re exempt.”
Congress would have to update Title 13 or pass explicit language authorizing AI-based privacy systems.
Right now, the Bureau’s lawyers interpret the law very conservatively — they prefer over-protection to the risk of an identification lawsuit.
🧩 3. The real opportunity — a “next-gen Census” prototype
In research circles, people are already working on this. A few examples:
The Harvard Privacy Tools Project and the MIT Trustworthy Data Initiative are designing AI frameworks for statistical confidentiality with verification.
The National Institute of Standards and Technology (NIST) has an active roadmap for “AI assurance in federal statistics” — which would let agencies use privacy-preserving computation under audited algorithms.
The UK Office for National Statistics (ONS) and the EU’s Eurostat are experimenting with synthetic census datasets built using generative AI, releasing realistic tables without exposing anyone.
So, yes — if the U.S. wanted to, it could run the original 2020 data through a secure AI model, generate a privacy-compliant synthetic dataset, and let states verify seat apportionment effects without ever seeing a single real person’s data.
That would take political will, cross-party trust, and updated law — not new technology.
🚀 4. The bottom line
You’re right on the money:
AI makes true transparency without privacy violation possible.
What’s missing is the legal framework and institutional courage to use it.
“...U.S. Code mandates that mid-decade census results “Shall not be used for apportionment of representatives in Congress … nor shall such information be used in prescribing congressional districts.”
Yes, this is ridiculous to allow to stand. That should be an easy tweak. When they have ADMITTED errors, it should be corrected as soon as the errors are identified.
If the data has not been released, how was it determined that there are 6 questionable seats ?
maybe it is 7, maybe it is 2. How does anyone know?
and if they do know, I would like to see their proof
“and if they do know, I would like to see their proof”
It is the Census Bureau itself that made the admission. Certainly THEY have the proof & no further investigation is needed. They should fix apportionment, as per their own data.
Didn’t President Biden change which agency
was in charge of collecting the 2020 census?
Or is my memory faulty?
We wouldn't want to be uppity with our Good Friends Across The Aisle, would we?
Nah!
” even the need for congressional/senatorial representation is no longer needed, since when it came to representative government, we needed to elect and send them off in order for each of them to represent the views and needs of each area they represented”
It wasn’t just time and distance that mandated representative government versus popular plebiscites.
It’s a concern that mob rule and emotional voting would lead to catastrophes.
A representative is supposed to provide a buffer to some of that.
But it’s still true that trekking to DC to be assimiliated by the Borg is no longer necessary. So reps can operate from home territory.
But instead of the out of date 435 congressional quota, there should be far more: we have lost representation because each congressman represents too many warm bodies. And in Democrat districts, cold bodies too.
So the size of the House should be expanded. In 1787, with a population of roughly 4M, each congressman represented about 9000 people. Maybe they were targeting 10K / rep as they knew growth would happen. So if we go back to that...we would need about 27,000 Congressmen assuming that there are really only about 270M actual citizens in this country. And it would mean that the districts would shrink dramatically. But this would actually make the reps closer to the people.
Such logistical hurdles would be a problem, but of course it can be managed now. And we wouldn’t need to build a giant amphitheater to accommodate them all in DC. They just need to log in and vote, or watch hearings / participate on line.
The Senate should keep its size since it represents the States. More importantly, the State legislatures should be appointing them again, reversing the mistakes of the “progressive” era.
Each state has its representative body, and each state has a governor.
If those people do represent the state, and the governor represents the wishes of the people at the time of the election, then, there is no need for a national congress or national senate. Let the in-state reps and governors express the states’ constituencies, directly. A national congress is just a duplication of what the in-state reps and governors do. Phone ‘it’ in, or send it in via internet, or if all else fails, send the governor to Washington to vote for the people of his state. It’s not months-long journey anymore to get the people’s wishes and wants expressed.
Only 6?
I’m trying to remember who is in charge of the “Executive Branch” ... just can’t quite think of his name ...I'm betting you'd like to/can't wait to forget.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.