An initiative of ARSIC Romania

H.E.A.R.T.

Human Ethics And Rights in Technology

An honest conversation on AI, humanity,
and the crisis of our world.

read the manifesto

Where we stand

Technology is part of our daily life. How it grows from here is still a human decision.

This is not a manifesto against technology. Tools have been with us from the beginning, extending our reach, our memory, our capacity to care for one another.

But technology is never neutral. It carries the priorities of the people who design it, fund it, regulate it and deploy it. A system that decides who gets a loan, a diagnosis, a place at a school or a second chance is not a technical object. It is a decision about a person, taken faster and at a larger scale than any institution before it.

So the question is not whether we want AI. The question is how we build it so that people stay at the centre.

Two directions

Technology that carries responsibility towards people
Decisions that can be explained to the person they affect
Consent that is asked for, not assumed
Errors that have an address, and someone who answers at it
Data that stays with the people it describes
Automation that gives back time instead of removing people
Technology without responsibility towards people
Decisions no one can explain, including the people who deployed them
Consent buried in documents written to be skipped
Failures that dissolve between vendor, client and algorithm
Data collected once and reused forever
Efficiency measured in people made redundant

The direction is chosen by the people who build and use them.

H.E.A.R.T.

Five commitments.

Human Ethics And Rights in Technology. The small word in the middle is not filler. It is the one that binds the others together: someone has to answer.

H
Human

The human before the system.

Dignity is not a feature that can be configured, downgraded or traded for speed. It does not depend on productivity, on a credit score or on a risk category. Any system that touches a human life inherits the obligation to treat that life as an end in itself, never only as an input.

01 A person must always be able to reach a person.
02 Everyone deserves to be treated as a person, not as a probability.
03 Vulnerability is a reason for more care.
E
Ethics

Respect for human values.

Ethics works best at the beginning: when the problem is framed, the data is chosen and the incentives are set. Placed there, it is not a brake on the work. It becomes part of the design, and it usually makes the result better.

01 Ask what the system does when it fails, before asking what it does when it works.
02 Test together with the people most affected, not only with the average user.
03 If a decision cannot be explained plainly, it is not ready to be taken automatically.
A
And — Accountability

Responsibility stays with people.

In a long chain of suppliers, models and data, responsibility easily loses its address. Keeping it simply means keeping that address visible: someone who knows the system, can explain it and can put it right. It helps the people who use the system, and it helps the people who built it just as much.

01 Every system in use has a person who knows it and answers questions about it.
02 Any automated decision can be reviewed, and asking for a review costs the person nothing.
03 An independent review makes a good system easier to trust.
R
Rights

Human rights are sacred, in any technology.

Privacy, non-discrimination, freedom of expression, access to justice, protection of children. These rights hold in an automated process exactly as they hold anywhere else. The law is already written. Our work is to make it easy to use, in a language and at a pace that actually help people.

01 The right to know that a machine was involved.
02 The right to a human review that is real, not procedural.
03 The right to a future that is not decided by your past.
T
Technology

Technology as support and human growth.

We can want abundance, better medicine, better schools, more time to live. And we can choose, among the paths that are technically possible, the ones worth taking. Progress is what leaves people better off, with someone standing behind it.

01 Build for the edge of attention, not for the centre of it.
02 Measure success in what people gain, not in what they cannot leave.
03 Slower is sometimes the more advanced engineering decision.

What we do

The principles that govern this project.

Monitoring

We follow how AI and automated systems are deployed in services that affect people, in Romania and in Europe, in the public sector and in the private one.

Audit

We examine systems, procedures and documentation against the obligations their operators already have, and we publish what we find.

Guidelines

We write practical guidance for organisations that want to use AI responsibly and do not know where to start: NGOs, schools, clinics, small companies, public institutions.

Education

We help people understand automated decisions, ask questions about them and, when needed, ask for a review, as part of the digital education programmes ARSIC already runs.

Official standing

We are working towards recognition as a trusted flagger under the Digital Services Act, through the procedure established in Romania by ANCOM Decision no. 336/2024. The status is granted to independent organisations with proven expertise in identifying illegal content, and it gives reports made in good faith a clear, priority path to the platforms.

empowering communities

Who we are

H.E.A.R.T. is an initiative of the Romanian Association for Community Support and Initiative.

ARSIC has worked with communities since 2007, through projects in education, civic participation and media. Digital and AI ethics is the newest of these programmes, and it grows out of the same work.

proinitiative.ro

We would rather be early than right too late.

If you are part of this world — building, using, teaching, regulating or simply living alongside these systems — this conversation is yours too. Bring your opinion. We will take it into account.

Join the conversation

References

The reading that shaped this text.

Leo XIV, Magnifica Humanitas (2026)

Encyclical letter on human dignity, work and artificial intelligence.

The Hume Initiative, Guidelines

Six principles for empathic AI: beneficence, emotional primacy, scientific legitimacy, inclusivity, transparency, consent.

Marc Andreessen, The Techno-Optimist Manifesto (2023)

An influential current of techno-optimism, useful to follow closely.

ANCOM, Notificatori de încredere

The Romanian procedure for trusted flagger status, Decision no. 336/2024.

Regulation (EU) 2022/2065 — Digital Services Act

The legal basis for trusted flaggers and for platform accountability in the European Union.

H.E.A.R.T. an initiative of ARSIC Romania
heart.ngo · 2026