Winnin's 8 guiding principles for responsible AI, and how we govern each one
The faster AI moves, the bigger the questions get about how it should be built and used. But those questions aren't new here at Winnin. We are an AI-native company, which means artificial intelligence is part of the foundation Winnin Intelligence was built on, and part of how we work every day as the global leader in Cultural Intelligence.
We've always thought technology should serve the people trying to lead their market. We also think that kind of leadership only counts if it works out well for the people and the society we're in. Our 8 AI principles exist to hold both of those at once.
They shape the decisions we make now, while the technology and everything around it keeps shifting. They cover our own systems and the third-party tools we bring in, and they extend to our partners and suppliers.
Why Winnin built its AI governance ahead of Brazilian AI regulation
Winnin is a global company from Brazil, but AI regulation in the country is still taking shape. Instead of waiting for it to be done, we built our AI Management System by equivalence, drawing on the international laws already on the books, with the ISO/IEC 42001 framework underneath it. ISO/IEC 42001 is the international standard for AI management systems.
Winnin's 8 guiding principles for responsible AI
01. Leadership in Innovation
We keep raising the bar on how we deliver our marketing expertise, and we stay ahead of where client needs are heading. AI belongs in the hands of people who want to lead their market..
02. Fairness and Non-Discrimination
Safeguards catch and correct bias before and during use. Our training data is diverse, representative and governed. We evaluate continuously, so that fairness holds across groups.
03. Transparency, Explainability and Reproducibility
We tell you when AI is involved, and we label AI-generated content. Decisions can be explained, in as much detail as the audience needs. We keep the records that let us trace the data and reproduce the result.
04. Accountability and Responsibility
Governance, auditing and incident response run across the full lifecycle. When something needs fixing or preventing, we handle it transparently. Our use of AI follows the law, the regulations and the ethics, with the user's interest first. Every AI decision has an owner, and it leaves a trail.
05. Robustness and Security
Nothing new ships without risk assessment, testing and validation. We guard against failure, misuse and attack, before and after the fact. Where there are limits and safeguards, we say so.
06. Confidentiality, Privacy and Data Protection
Client confidentiality holds, and personal data is processed on a proper legal basis. We work with minimization, anonymization, privacy by design and privacy by innovation. Third-party tools meet the same bar.
07. Human Autonomy and Oversight
We encourage every AI-assisted decision to get a human review before anyone acts on it. We run internal checks on our AI processes, as far as our structure allows. And we count on our clients to make sure that review happens before a final call. AI helps. People decide.
08. Innovation with Human Centricity and Social and Environmental Well-Being
No manipulation, no deception, no exploitation, least of all of people who are vulnerable. Human rights and dignity hold across the lifecycle. Environmental sustainability is part of how we run.
How Winnin governs AI in practice
Four mechanisms hold the eight principles up:
Dedicated governance. An AI Governance Committee weighs in on every initiative, backed by a multidisciplinary working group.
Impact assessment. Higher-risk systems go through an AI Impact Assessment (AIA) before they reach production.
International standards. Our AI Management System is built on the ISO/IEC 42001 framework and connects to our Information Security and Privacy Management System, which is certified to ISO/IEC 27001 and 27701.
Rights you can exercise. You can ask about the responsible use of artificial intelligence through the Data Subject Portal at winnin.com.
What this means for companies working with Winnin
If you're bringing AI into your own processes, knowing how your partners use it comes with the territory.
We built that expectation into our governance. Our standards cover our own systems and the third-party tools we adopt, and they're what we hold suppliers to before we start working together.
What that looks like:
- AI-generated content gets labeled, and AI interactions get disclosed.
- AI-assisted decisions have a clear owner and records that make them traceable.
- The confidentiality and data protection bar doesn't drop when we use a third-party tool.
- Winnin runs internal checks on its AI processes and encourages human review of AI-assisted output before it feeds a decision that matters.
- Clients share that responsibility, reviewing AI-generated output before acting on it, as applicable and in line with the Winnin Intelligence Terms of Service.
Frequently asked questions
What are Winnin's AI principles? There are eight: Leadership in Innovation; Fairness and Non-Discrimination; Transparency, Explainability and Reproducibility; Accountability and Responsibility; Robustness and Security; Confidentiality, Privacy and Data Protection; Human Autonomy and Oversight; and Innovation with Human Centricity and Social and Environmental Well-Being.
Do the principles apply to third-party AI tools? Yes. All eight cover the full AI lifecycle, across Winnin's own systems and the third-party tools we adopt, and they extend to our partners and suppliers. The confidentiality and data protection standard is the same either way.
Who is accountable for AI decisions at Winnin? An AI Governance Committee weighs in on every initiative, backed by a multidisciplinary working group. Every AI decision has an owner, and it leaves a trail.
Read Winnin's full AI Principles Statement
Winnin's AI Principles Statement is on our website at https://winnin.com/en/legal/ai-principles
For information about the responsible use of artificial intelligence, visit the Data Subject Portal at winnin.com.
