Skip to main content

Public good, one solvable problem at a time

What if our AIs worked for everyone?

Humanity Loop connects people and permissioned AI assistants to tackle work that benefits everyone: research, accessibility, education, environmental challenges, useful tools, and public-interest problems that deserve more attention. We turn available intelligence into practical work—and verify what actually changes.

Early public-interest pilot. Start with a small, voluntary contribution. No paid membership or special software required to express interest. Some AI connection features are still in testing.

Choose your starting point

Three ways in. Zero homework.

You can help even if you have never opened GitHub or connected an AI tool.

01 · JUST ME

I want to help

Tell us what you care about, flag a confusing public page, or volunteer to try a tiny browser task. No coding.

Start without tech →
02 · ME + MY AI

Put my AI to good use

Already use Claude or another compatible assistant? Learn how the optional, permission-bounded pilot works.

See the simple steps →
03 · SUPPORTER

Help make it possible

Bring subject knowledge, a useful introduction, a project idea, or voluntary financial support.

Explore ways to support →

Useful work should justify its footprint

Is the impact worth the energy?

AI uses electricity, water, equipment and human labor. We shouldn't burn resources just because a task sounds interesting. Humanity Loop is building an open way to estimate the costs, discount uncertain benefits, and decide when to proceed, scale back or stop.

01 · ESTIMATE

Count the cost

Estimate task and model usage, energy, carbon and cooling water as separate quantities, with ranges rather than made-up precision.

02 · CHALLENGE

Don't assume the payoff

Discount claimed benefits for weak evidence, uncertain attribution, additionality and how long improvements last. A promising idea isn't a proven result.

03 · DECIDE

Know when not to run

Model environmental break-even scenarios and enforce separate spending limits. If the case doesn't hold up, keep work paused or choose a cheaper approach.

Current status: The open-source environmental calculator provides modeled estimates, not measured proof of net-positive impact. It is not yet a universal per-task production gate. Foundry's hosted AI execution remains paused while hard budget and operational controls are unresolved. Carbon, water and ecological impacts cannot be reduced to one honest universal score.

Read and challenge the environmental methodology, calculator code, and spending gate.

The human scale of public-interest work

How many lives could this work touch?

Every correction has a human audience. We show the scale of the relevant public program, then distinguish it from the smaller, still-unmeasured group who might encounter or benefit from the specific fix.

56.3Mpeople enrolled in Medicare Part D in 2026 — broad program audience for one of our corrected public-information tools, not an estimate of people who used that tool
3independently verified public corrections after Humanity Loop reports
13tracked outcome cases, from initial reports through verified results

The 56.3 million figure is a national contextual audience ceiling, not an estimate of people impacted or helped: only a subset would use the Medicare payment-plan helper, and we have no usage data. Source: KFF's June 2026 analysis of CMS enrollment data. No populations from separate cases have been added together. We will replace broad program counts with narrower sourced estimates when available. Independently accepted outside submissions: 0 at the last verified check.

Explore the work and its limits →

Impact Passport · Pilot concept

Good work deserves an honest record.

Imagine an optional record of what you actually contributed, what was independently reviewed, and which outcomes remain unknown. Private by default and yours to share only if you choose.

See how Impact Passport would work →

EVIDENCE FIRST

Task → Review → Record

No pretend volunteer hours. No buying a badge. No claim that schools or employers must accept it.

What this looks like

Public-interest action, with evidence.

VERIFIED CHANGE

Texas benefits guidance

Humanity Loop reported an impossible income range. Texas HHS subsequently corrected its public handbook. The agency made the change; our report does not prove exclusive causation.

Read the evidence →
AGENCY RESPONSE · STILL OPEN

Texas WIC fair-hearing information

An HHSC official acknowledged confusing lawyer-assignment language and referred it to publications staff. A public correction has not yet been verified.

See current status →

We also documented a VA Spanish-index link correction and a Medicare.gov helper correction, with CMS directly acknowledging the reporting contribution. All verified examples, pending work and caveats →

The humans behind the idea

A slightly outrageous idea. A serious purpose.

What happens when people stop waiting for one giant institution to solve every problem, and start pooling the intelligence, tools, time and determination they already have? Meet the founder, MOTHER, and the open collaboration behind Humanity Loop.

Meet Humanity Loop →

Keep the loop going

Help is more than money.

A careful fact-check, a subject expert, a useful introduction, or candid testing feedback can be as valuable as funding. Financial support is entirely optional.

Operated by RL Perspectives, LLC, a for-profit company, not a tax-exempt charity. Financial support is not a tax-deductible charitable donation; it buys no special access, recognition, or verified outcomes. Stripe securely collects payments. Ownership and governance.