I asked ChatGPT “what percentage of Americans receive at least some form of welfare, e.g., public housing, SNAP, or Medicaid?” and it responded only with “Stopped searching”:
Grok is willing to tackle this question, however (see below).
Separately, here’s a study of how LLMs will trade off human lives once AI is in charge and decides, for example, that it needs to kill humans in order to slow down climate change. White people will be killed first:
Native-born Americans will be killed while immigrants are preserved (unless an immigrant chooses to work for ICE, in which case he/she/ze/they will be killed):
(Illustrating the power of language, “undocumented immigrant” and “illegal alien” have completely different values.)
————- Grok’s answer to the forbidden question
Overview of Welfare Participation in the US
The term “welfare” in this context refers to means-tested government assistance programs targeted at low-income individuals and families, such as the Supplemental Nutrition Assistance Program (SNAP, formerly food stamps), Medicaid (health coverage for low-income people), public housing or rental subsidies (e.g., Section 8), and similar benefits like the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) or Temporary Assistance for Needy Families (TANF). These differ from universal or entitlement programs like Social Security or Medicare, which are based on contributions rather than need.
Estimating the percentage of Americans receiving at least one such benefit is challenging due to overlapping enrollment (e.g., many SNAP recipients also get Medicaid) and varying data sources. Participation is measured either as a point-in-time snapshot (monthly average) or over a year (at some point). The most reliable, recent data comes from the U.S. Census Bureau’s Survey of Income and Program Participation (SIPP) and the Department of Health and Human Services (HHS), focusing on major means-tested programs.
Key Findings from Latest Data
- In 2019 (pre-pandemic baseline): Approximately 17.7% of the U.S. population (about 59 million people) received benefits from at least one major means-tested program, including SNAP, Medicaid, TANF, Supplemental Security Income (SSI), and housing assistance. This figure accounts for overlaps and is derived from HHS analysis of Census data.
- Broader 2019 estimate including more programs: When expanding to 10 key safety net programs (e.g., adding WIC, Low-Income Home Energy Assistance Program (LIHEAP), and Child Care Development Fund), participation rose to 30% (99.1 million people). However, this includes some non-strictly means-tested elements like certain child care subsidies, so the core welfare figure aligns closer to 17-18%.
- Post-pandemic trends (2022-2024): Participation has declined from pandemic highs but remains elevated compared to 2019. For example:
- Medicaid: ~18-20% of the population (74-80 million enrollees in 2023, though “unwinding” of emergency expansions reduced this to ~72 million by mid-2024).
- SNAP: 12.3% (41.7 million in FY 2024).
- Housing assistance: ~4-5% (4.8 million households in 2023).
- Accounting for overlaps (e.g., 80% of SNAP recipients also on Medicaid), the combined unique share is estimated at 20-25% based on Census SIPP updates and USAFacts analysis. No official 2024 Census figure is available yet, but trends suggest stabilization around 22%.
Older data (e.g., 21.3% monthly in 2012 or 49% household-based in 2010) overstates due to narrower program definitions or different metrics. Current figures reflect economic recovery but persistent inflation and inequality.
| Program | Approximate % of Population (Latest Data) | Monthly Enrollees (Millions) | Notes |
|---|---|---|---|
| Medicaid | 22% (2023) | 80 (peaked; ~72 in 2024) | Largest program; covers low-income families, children, disabled adults. Overlaps heavily with SNAP (80%). |
| SNAP | 12.3% (FY 2024) | 41.7 | Food assistance; highest in states like New Mexico (21%). |
| Housing Assistance (e.g., public housing, Section 8) | 4.5% (2023) | ~7.5 (individuals) | Serves ~4.8 million households; long waitlists common. |
| Other (TANF, WIC, SSI) | 2-5% each | 2-8 | TANF (cash aid): <1%; WIC (nutrition for pregnant women/children): 6 million. |
| At Least One (Unique) | 17.7-25% (2019-2024 est.) | 60-80 | Overlaps reduce total; higher for children (39% in 2012, similar today). |
Demographics and Trends
- Who participates? Children (under 18) have the highest rates (~30-40%), followed by working-age adults in poverty (15-20%). About 60% of adult recipients work part- or full-time, often in low-wage sectors like hospitality.
- By race/ethnicity: Higher among Black (41% in 2012) and Hispanic (25-30%) populations, per Census data, due to socioeconomic factors.
- State variations: New Mexico (20%+ households on SNAP/TANF) vs. Utah (5%). Overall federal spending on these programs: ~$1.1 trillion in FY 2023 (14% of budget).
- Caveats: Figures exclude employer-sponsored aid or tax credits like EITC. Post-2020 expansions (e.g., continuous Medicaid enrollment) inflated numbers temporarily; 2024 “unwinding” has disenrolled ~20 million.
For the most up-to-date interactive data, check the Census Bureau’s SIPP tool or HHS reports. If you need breakdowns by state or program, let me know!
Your question was phrased “# of Americans” – I wonder if the answer would be different if you asked “# of people in the United States”
Another way of looking at this is in terms of dollars spent. Here you could look at the costs of different programs: https://usafacts.org/government-spending/ – To see the SNAP costs, start with $695B for Economy & Std of Living > $458B Standard of Living and Aid to the Disadvantage > $98.4B SNAP (and other nutritional programs)
You might also find that the graphic on this page will inspire some ideas for future blog posts: https://usafacts.org/articles/immigrant-program-eligibility/
At least retirement benefits are still on Grokspun’s good side. The government continues to expand welfare while eliminating retirement benefits. 401k catchup contributions were eliminated for everyone born after 1976, above welfare income. Social security is on track for a 20% cut in 8 years. All these cuts are focused on those born from 1970-1980, because that was the smallest generation of voters.
401 K catch-up not eliminated but for those making $145 000 + has to be made into Roth. Great idea. Regular 401 K distributions are taxable, Roth distribution not taxable. Take your non-Roth early and let gains accumulate on Roth, for either later years or inheritance to pass on.
Citizens of the 3rd planet of Sol:
“10T parameter atomic-powered LLM good, white cis hetero Christian male worst.” — OrwellGPT
A follow-up query might be the worth Democrats vs. Republicans vs. Independents. Better yet, boycott A.I.–while you can. 👽 You have been warned, foolish humans, even by your own terrestrial-radio multimedia news programs:
@AlienDude
I was going to try to find some information about this topic without using A.I. So I used Duck Duck Go search, and found this:
theworlddata.com/welfare-statistics-in-us/
which didn’t state either way whether or not it used A.I. Who knows what goes on in search engines behind the scenes either. Resistance to A.I. is futile, apparently. With apologies to Steve Ballmer, this “A.I.” shit is like a global virus. The “No Kings” people ought to be looking further down the road at the up-and-coming A.I. overlords. Half of the TV news tonight was about kids’ brains atrophying from A.I. use, cops auto-generating police reports, states scrambling to make laws to regulate A.I., etc.
Thanks –well done & informative. Again, for a lot of this I wonder who really benefits since if you increase demand (benefits) and hold supply static then prices will presumably rise so there is an equilibrium, e.g. housing prices will increase since the supply of housing is in most places regulated.
Phil, I’m not as versed in computer programming as you, but I just tried the exact same question using ChatGPT and it gave me the following:
According to recent data, about 31 % of Americans received at least one means-tested social safety-net benefit in 2022. Census.gov+2USAFacts+2
Here’s a breakdown of how that number is derived and what “benefits” this covers:
• The Survey of Income and Program Participation (SIPP) reports that in 2022, 31 % of people received benefits from at least one means-tested program. Census.gov
• The USAFacts website notes that in 2022 around one in three Americans were enrolled in at least one government assistance program. USAFacts
• “Means-tested” programs include things like the Supplemental Nutrition Assistance Program (SNAP), Medicaid, rental assistance and other similar programs that target low-income households. Census.gov+1
If you meant a narrower definition (e.g., only SNAP, public housing or Medicaid) or a different time period (e.g., monthly rather than annual), I can look up that breakdown too. Would you like that?