DLYog Lab Research Services LLC · Version 1.0 · Last updated:
Preamble: AI for the good of people
We believe the purpose of AI is to advance human well-being. Its value comes from helping people live better, make informed choices, and pursue what matters to them. We want small businesses to succeed and become AI native in ways that serve their owners, employees, customers, and communities.
Our founders’ conviction is simple: progress should preserve human dignity, agency, joy, and connection. Generative AI can automate many activities. That capability does not, by itself, make automation desirable. We will not automate away experiences people enjoy or unnecessarily diminish what makes those experiences meaningful.
This constitution expresses the principles behind our internal AI commitments and the boundaries that guide the work we accept, design, recommend, and operate. It applies to our AI services, consulting, and automation work.
Our governing principles
Each principle connects a belief to a practical commitment. Human benefit is the purpose against which we assess the work.
On smaller screens, scroll horizontally to read all three columns.
| Principle | Why it matters | Our commitment in practice |
|---|---|---|
| I. Human well-being | AI should improve people’s lives and contribute to the good of people. | We assess who benefits, who may be harmed, and whether automation serves a meaningful human need. |
| II. Meaningful human experiences | Enjoyment, participation, choice, and connection have value beyond efficiency. | We preserve activities people value and decline work that unnecessarily replaces those experiences. We categorically refuse automated checkout or purchasing on a person’s behalf. |
| III. Human agency and oversight | People should understand and retain control over consequential actions. | We define human review, approval, and ways to pause permitted automation. Customers choose whether to adopt permitted workflows; we remain responsible for our own work. |
| IV. Respect for permission | Being able to perform an action does not establish the right to perform it. | We respect applicable laws, third-party terms, AI usage policies, and automation restrictions. We do not bypass access controls or anti-bot safeguards to carry out prohibited work. |
| V. Prevention of harm | Business goals must be pursued with care for everyone affected. | We decline fraud, deceptive impersonation, harassment, unauthorized access or data collection, unlawful discrimination, and misuse of personal information. |
| VI. Privacy and security | People and businesses entrust us with information and access. | We limit data and permissions to what a workflow needs, protect credentials, and review provider data handling before sharing information. |
| VII. Honesty about AI | Informed decisions depend on understanding uncertainty and limitations. | We explain AI’s role, disclose known limitations, and avoid presenting uncertain outputs as verified facts. Important results need verification. |
| VIII. Fairness and care | Errors and harmful bias can affect people differently. | We test relevant scenarios, examine errors and unfair outcomes, and choose safeguards proportionate to the workflow’s impact. |
| IX. Informed small-business adoption | Small businesses deserve understandable guidance about benefits and risks. | We explain tradeoffs, data needs, failure modes, and review requirements in plain language so customers can make informed choices. |
| X. Accountability and improvement | A commitment requires action when a concern or failure arises. | We review reported concerns promptly, take appropriate corrective action, and revisit workflows and this constitution as we learn. |
A firm boundary: checkout and purchasing
Under our AI constitution, we will not build or operate automated checkout or purchasing services that buy on a person’s behalf, even where a platform permits them. Purchasing decisions and the act of completing a purchase remain with the person.
This boundary reflects our founders’ belief that shopping can be a valued human experience. Customers and individuals we have spoken with have described enjoying it. Those conversations inform our position; we recognize that preferences differ. Our choice is to preserve human participation rather than pursue automation solely because it is possible.
How we apply this constitution
- Start with the human purpose. Identify the benefit, the people affected, and the experiences or choices the workflow could displace.
- Check the boundaries. Review our constitution, applicable requirements, and platform permissions. Customer consent or commercial opportunity cannot override our ethical boundaries or a third party’s restrictions.
- Explain the risks and controls. Discuss inaccurate or fabricated outputs, harmful bias, privacy exposure, security threats, unintended agent actions, and service or policy changes. Agree on data access, testing, review, and approval appropriate to the impact.
- Proceed only with a permitted approach. When permission or safety is unclear, pause the affected work and seek clarification. Where possible, suggest an authorized integration or a workflow that preserves human participation.
- Review and respond. Revisit workflows as conditions change. Pause or correct affected automation when problems warrant it.
We respectfully ask customers to make requests consistent with these principles. We will decline or revise requests outside them. Customers make the final adoption decision for permitted workflows, and decisions requiring professional judgment need an appropriately qualified person. AI can still make mistakes despite safeguards.
Questions, concerns, and accountability
To ask about our internal AI constitution or report a potential responsible AI violation involving our work, email info@dlyog.com.
For a concern, describe the service or workflow, what happened, and relevant dates or links. Please do not send passwords, access tokens, or sensitive personal information.
We will promptly review concerns, investigate as appropriate, and take corrective action where needed. This may include pausing affected work, correcting a workflow, or coordinating with the relevant customer or provider. We remain accountable for our own actions.
Review and public references
This constitution records our current commitments. We will revisit it as technology, risks, and our understanding evolve, and identify revisions by version and date.
Our principles express DLYog Lab’s own values. Public guidance consulted includes the NIST AI Risk Management Framework on risk management and the OECD AI Principles on human rights, transparency, safety, and accountability. These references do not imply certification or endorsement.
Read this constitution alongside our Privacy Policy, Transparency Policy, and Terms & Conditions.