This site is under construction - will be ready soon.
Your Guide to Our Vision

AI · Healthcare · Hiring · Accountability

The Right to a Human

What the Federal Government Walked Away From, and What We Will Put Back

By Vinnie CordovaCordova 2028: Moving Forward Together
The Right to a Human: a woman, nurses, and an older man look toward the light beside a screen reading AI decision, application rejected. The right to know, the right to reach a person, the right to hold someone accountable.
Campaign design team

When an AI system can decide whether you get care or get hired, you have the right to know it, the right to reach a person, and the right to hold someone accountable.

In When the Human Disappears, I asked a question: what does it tell us about an institution when it decides that a machine is enough to represent its relationship with another human being?

This piece answers the question that follows from it.

What are we going to do about it?

The questions alone are no longer enough. While hospitals and employers have been rushing to put AI between themselves and the people they serve, the federal government has been quietly backing away from protecting those people.

I want to be clear about what happened, and clear about what I will do differently.

While you weren’t looking

What Happened

In January 2025, the Equal Employment Opportunity Commission removed its guidance on artificial intelligence and employment discrimination from its website (source).

That guidance told employers something simple: if your hiring software screens people out unfairly, you are responsible. It explained how automated tools can discriminate against applicants by race, sex, age, or disability, even when no one intended it. It told employers to test their tools. It told applicants with disabilities they could ask for an accommodation.

It was taken down after executive orders rescinded the prior administration’s AI policies.

Then it went further. In September 2025, the EEOC directed its field offices to close investigations into disparate impact discrimination (source). That is the kind of discrimination automated hiring systems are most likely to produce: a “neutral” rule that quietly shuts out a whole group of people.

AI Decision

☑ Screening: Automated

☑ Review: AI System

☑ Human Review: None

☑ Federal investigation of disparate impact: Closed

Application: Rejected

Read that again.

The type of harm an algorithm is most likely to cause is the type of harm the federal agency responsible for stopping it chose to stop investigating.

The laws themselves did not change. Title VII, the Americans with Disabilities Act, and the Age Discrimination in Employment Act still apply when a machine makes the decision (source). An employer still cannot discriminate through software that it could not lawfully do by hand.

But a law no one enforces protects no one in practice.

The rules did not get weaker. They got quieter. And quiet rules protect the powerful.

Your dignity should not depend on your zip code

The States Are Filling the Gap

Some states stepped in.

Here in California, new regulations under the Fair Employment and Housing Act took effect in October 2025. They apply anti-discrimination law directly to automated decision systems used in employment. Other states have passed their own laws, each with a different standard (source).

The courts are moving too. In Mobley v. Workday, a case in federal court in California, an applicant argued that the software vendor itself, not just the employer, can be responsible when an AI screening tool rejects people unlawfully. That case has been allowed to move forward on behalf of a larger group of applicants.

That is progress. But it is not a system.

A job seeker in one state should not have fewer rights than a job seeker in the next state over simply because their legislature moved slower. A patient should not have to research their state’s AI laws before calling a hospital.

Your dignity should not depend on your zip code.

It is not just a chatbot

Why This Matters More Than It Sounds

I have heard the argument that this is a minor issue. That a chatbot scheduling an appointment is not a civil rights question.

Some of it is minor. An appointment reminder is not a threat to anyone.

But that is not where this stops.

AI systems are now screening résumés, scoring recorded video interviews, ranking applicants, and deciding who advances before any human being looks at them. AI voices with human names are answering hospital phone lines. Automated systems are sorting patients, handling billing disputes, and deciding who gets routed to a person and who does not.

If success means

Fewer staff and lower costs

that is what the system will deliver.

If success means

Better care and fairer hiring

the same technology can deliver that instead.

Every one of those systems was built by someone, bought by someone, and deployed by someone who decided what success means.

The tool follows the priorities built around it. Right now, too many of those priorities are being set with no public accountability at all.

Private equity and profit before people

Who Is Holding the Tool

I said in The Tool Reflects the Nation that a tool does not decide its purpose. The person holding it does.

So we have to talk honestly about who is reaching for this tool first, and why.

It is private equity firms. It is institutional holders demanding higher returns every quarter. It is the owners and executives who believe a balance sheet matters more than the people it is supposed to serve.

Private equity is not built the way human life is built.

What a human being needs

Stability. Care. Time. Someone who stays.

What the model is built to do

Buy → load with debt → cut costs → extract cash → sell

usually within a few years

Its purpose is the return to the investor. Not the patient. Not the worker. Not the community.

That is not an insult. It is the model, and we already know what it does when it touches healthcare.

It has already been tested on human beings. Here are the results.

Hospitals

+25.4%

hospital-acquired conditions

A Harvard-led study in JAMA examined hospitals before and after private equity bought them. After acquisition, Medicare patients suffered a 25.4% increase in hospital-acquired conditions, the preventable harms hospitals are supposed to stop. Falls rose. Bloodstream infections from central lines rose. Surgical site infections doubled, even though these hospitals were performing fewer surgeries. The researchers pointed to reduced clinical staffing. (source)

Patient experience

↓

ratings and staff responsiveness

A second JAMA study, published in 2025, found that patients’ ratings of private equity–acquired hospitals declined after the buyout, including how responsive the staff were, and kept getting worse over time. (source)

Nursing homes

20,000+

estimated lives lost, 2005–2017

A National Bureau of Economic Research study found that private equity ownership of nursing homes increased the short-term death rate of Medicare patients by about 10%. Frontline staffing went down. Taxpayer spending per patient stay went up 11%. Interest payments rose by about 325%, and lease payments rose by 75%. The money was going somewhere. It was not going to the residents. (source)

Steward Health Care

$800M

taken out by Cerberus

In 2010, private equity firm Cerberus Capital Management bought a chain of nonprofit Catholic hospitals in Massachusetts and renamed it Steward. In 2016, Steward sold the land and buildings under its hospitals to a real estate trust for $1.25 billion and was left paying rent on hospitals it used to own. Cerberus walked away having made roughly $800 million. In May 2024, the whole system, by then 31 hospitals in eight states, collapsed into bankruptcy. Hospitals closed. Researchers documented more than 650 instances of deficient care across Steward hospitals. Cerberus denies wrongdoing. Patients and communities were left with the consequences. (source)

As of early 2024, at least 460 hospitals in this country were owned by private equity (source).

Now put artificial intelligence in those hands.

The private equity playbook already treats staff as the first cost to cut. AI gives it a faster, quieter way to do it. Replace the nurse who answers the phone. Replace the recruiter who meets the applicant. Replace the person who could have noticed that something was wrong. Then call it innovation.

In the hands of people who see humans as a cost, AI is not a tool for care. It is a tool for extraction at scale.

The industry will say it rescues struggling hospitals and nursing homes that would otherwise close. Sometimes capital does help. But the researchers behind these studies noted that most private equity buyouts target operations that are already successful (source), and the outcome data speaks for itself. And not every investor behaves the same way. That is exactly why we should judge them by results, in public, with the evidence on the table.

Here is what I need every person reading this to understand.

You are the value.

Not the debt they placed on the hospital. Not the real estate they sold out from under it. Not the quarterly return. You. Your work. Your care. Your life.

Every hospital, every company, every fund depends on human beings: the nurses, the patients, the workers, the families, the taxpayers. Without people, there is nothing to own and nothing to extract.

They need you to believe otherwise. They need you to believe you are a cost, a line item, a replaceable input. Because if you believe that, you will accept being treated that way.

That belief serves their essence, not yours.

I reject it. And I am asking you to reject it with me.

No vague promises

The Commitments

I do not believe in vague promises. People have heard enough of them.

So here is exactly what my administration will fight for, as part of the American Worker Protection and Human-AI Collaboration Act and through every executive authority available. These apply to what I call high-stakes uses: any AI system that affects a person’s:

EmploymentHealthcareHousingCreditInsuranceGovernment services
01

The Right to Know

If you are interacting with artificial intelligence in a high-stakes setting, you will be told, at the start, in plain language, before you share anything.

  • No AI system may present itself as a human being.
  • No AI system may use a human name, human persona, or digital face unless it clearly identifies itself as artificial intelligence at the start of the interaction and again whenever you ask.
  • “We mentioned it in the terms of service” does not count. Disclosure must happen in the interaction itself.

If an institution is proud of its technology, it should have no problem naming it.

02

The Right to Reach a Human

In any high-stakes interaction, you will have a clear, one-step way to reach a qualified human being.

  • The path to a human must be available at every stage, not buried behind menus.
  • “Escalation” to another automated system is not escalation.
  • The human you reach must have the authority to actually resolve your issue, not just read you the same script.
  • Institutions must publish how long it actually takes to reach a person, so the public can compare.
03

No Automated Rejection Without Real Human Review

No person will be denied a job, denied care, or denied a service by an AI system alone. Before any adverse decision becomes final, a human being must review it, and that review must be real.

  • The reviewer must have the time, the information, the training, and the authority to overrule the system.
  • Institutions must track how often reviewers actually disagree with the AI. A reviewer who approves 100% of automated rejections is not oversight. It is a signature on a machine’s decision.

“Human in the loop” will mean a human with power, not a human for show.

04

The Right to an Explanation and an Alternative

If an AI system influenced a decision against you, you have the right to know that AI was used, what it evaluated, the main reasons for the outcome, and how to challenge it.

  • If you have a disability, limited access to technology, or any circumstance that makes an automated assessment unfair to you, you have the right to request an alternative process.
  • That request can never count against you.
05

Responsibility Cannot Be Outsourced

An employer cannot outsource its legal or moral responsibilities to an algorithm. Neither can a hospital.

  • The institution that deploys an AI system remains fully responsible for its outcomes.
  • The vendor that builds and sells it shares that responsibility when its product causes harm.
  • No contract, terms-of-service clause, or indemnity agreement can waive a person’s right to challenge a decision or hold someone accountable for it.
06

Public Accountability Reports

Every institution using AI in high-stakes decisions will publish, in plain language:

  • Identification: where and when people interact with AI rather than a human.
  • Purpose: what each system does and whether it assists or replaces human work.
  • Decision authority: whether the AI informs, recommends, or decides.
  • Human access: how to reach a person and how long it actually takes.
  • Accountability: who is responsible when it gets something wrong.
  • Privacy: what data is collected, kept, shared, or used to train systems.
  • Fairness testing: how the system has been tested for errors, accessibility, and discriminatory effects, and the results.
  • Public benefit: what measurable improvement the technology has produced for the people it serves.

These will not be voluntary. Voluntary transparency is how we got here.

07

No Extraction Through Automation

Where private equity or other investment owners control a hospital, nursing home, or other essential care provider:

  • Ownership must be disclosed. Patients have the right to know who actually owns the place they are trusting with their lives, including every fund and holding company in the chain.
  • AI cannot be used to cut care below safe levels. No institution may replace clinical or patient-facing staff with automated systems if doing so reduces staffing below safe care standards.
  • Savings from automation must be shown. Any AI-driven staff reduction must be reported publicly, along with where the savings went: back into care, or out to investors.
  • The playbook gets scrutiny before the damage, not after. Acquisitions of essential care providers, sale-leasebacks of hospital property, and debt-funded dividend payouts will face public review before they close.
08

Restore Federal Enforcement

My administration will restore federal guidance on AI and employment discrimination and direct enforcement agencies to investigate automated discrimination, including disparate impact, with the seriousness it deserves.

  • The law already says you cannot discriminate. The government’s job is to make sure that remains true when the discrimination comes from software.

Who receives the value?

Where Do the Savings Go?

There is one more question I will not let disappear.

AI can cut an institution’s costs. That is real. But cost reduction is not the same thing as progress.

When a hospital eliminates thousands of hours of administrative work, does it hire more nurses? Shorten wait times? Lower the cost of care? Or does it keep the savings and charge the same?

When an employer cuts its recruiting costs with automation, does it pay workers more? Invest in training? Or does it simply reject more people with fewer staff?

The value created by artificial intelligence was built on generations of human knowledge, labor, and public investment. It should not flow only to owners and investors while the people whose work made it possible are told they are now too expensive to keep.

That is why the Act also protects workers’ right to have a voice when AI changes their jobs, to receive training and fair transitions, and to share in the productivity gains their work helps create.

The future should not belong only to those who own the machines.

You are not powerless

Until Then: What You Can Do Now

These protections will take a fight to win. Until then, you are not powerless.

With a hospital or medical provider

  • Ask whether you are speaking with an AI system or a person.
  • Ask to be transferred to a qualified human when your situation requires one.
  • Ask whether AI-generated information about your care has been reviewed by a medical professional.
  • Never delay urgent or emergency care because a provider uses AI. Your health comes first.

When applying for a job

  • Ask whether AI is used to screen, interview, or score applicants.
  • Ask whether a human reviews decisions before applicants are rejected.
  • If you have a disability, ask for an accommodation or alternative assessment. Federal law still protects that right.
  • If you live in California, know that state regulations now apply anti-discrimination law to automated hiring systems.

You are evaluating the institution just as much as it is evaluating you.

The measure of this technology

The Standard

I am not against artificial intelligence. I want every person in this country to have access to tools that expand what they can do.

I want doctors to have the best technology available, and more time with their patients because of it.

I want employers to find talent they would have missed.

I want people with disabilities to have technology that removes barriers instead of creating new ones.

But I will not accept a country where people have to fight their way past a machine just to be recognized as a human being. Where a frightened patient cannot reach a person. Where a qualified applicant is rejected by software no one is responsible for. Where the federal government watches all of this and decides to look away.

Did people receive better care?

Were applicants treated fairly?

Did employees benefit?

Was accountability preserved?

The measure of this technology will not be how many people it removes from the equation.

It will be whether life got better for the people affected by it.

Artificial intelligence is a tool.

Humanity remains the purpose.

And the right to a human being should never be a luxury.

I am a piece of the collective essence of this world that never gives up and needs every other piece to be whole.

Vinnie Cordova
Cordova 2028: Moving Forward Together
A government that listens, and proves it with results.
cordova2028.com

Community Comments

Community Comments

Share a public response to this post. Submissions are reviewed before they appear.

0 approved comments

Loading comments...

Comments are moderated for spam, abuse, and off-topic submissions.

Your age, area, and IP address are collected for moderation and internal reporting only.