In short

• The new Insight Association’s AI Code of Standards (2025) sets global benchmarks for ethics in research
• AI cannot replace human judgment, bias checks and oversight are mandatory
• Transparency on AI use, model type, and data sources is now required
• Participants’ privacy and consent are non-negotiable, especially for children & vulnerable groups
• Researchers who adapt will build more trust, richer insights, and stronger client relationships

Why this matters now

Market research is undergoing its biggest shift since the birth of online surveys. AI isn’t a buzzword anymore-it’s infrastructure. But with power comes responsibility, and the Insights Association’s 2025 Code of Standards is a wake-up call: we can’t innovate without trust.

These standards are about making sure it runs on the fuel of research itself: credibility, consent, and care.

The four pillars of ethical AI research

The Code distills our profession into four deceptively simple principles:

1. Respect people – participants are not datapoints; they’re humans with rights.
2. Transparency first – collect data openly, with informed consent.
3. Integrity always – no hidden agendas, no blending research with marketing.
4. Compliance everywhere – local laws + global ethical benchmarks.

In short: if a participant can’t trust you, your insights are worthless.

The AI rules that change the game

Here’s where things get specific for researchers using AI:

• No personal data in training sets: anonymization is the baseline
• Full disclosure: explain the purpose of AI use, model type (open vs. proprietary), and whether data is predictive, generative, or synthetic
• Humans stay in the loop: AI tools can scale, but human instruction and oversight keeps it ethical
• Bias checks: models must be audited for demographic and cultural bias, continuously, not once

At Glaut, we’re actively working with researchers worldwide on research-on-research to validate and refine AIMIs as a methodology. The first comparative study with G. M. Occhipinti suggested striking advantages over surveys: +129% more words per response, -53.6% less gibberish, and a +56% higher completion rate.

Imagine telling your client: “our methodology doesn’t just use AI; it meets the highest global standards for ethical AI research.” That’s authority, trust, and differentiation in one.

Protecting participants, protecting trust

The Code doubles down on duty of care:

• Consent must be explicit, never buried in fine print.
• Children and vulnerable populations require verifiable guardian approval.
• Participants have the right to withdraw, refuse, or access their own data at any point.

Glaut AIMIs are built with this in mind. Safeguards like voice-only mode, uncooperative detection, consistency checks, and interpretative scoring prevent fraud and filter out poor-quality responses in real time. In other words, data integrity is protected before it ever enters your dataset.

What this means for your next project

If you’re running research in 2025, the researchers who thrive will be the ones who can say:

• We use AI, but never at the cost of participant trust.
• Our methods scale like surveys, but probe like interviews.
• Every step, from brief to report, is transparent, customizable, and under your control.

Sound familiar? That’s exactly the promise of Glaut’s AI-moderated interviews (AIMIs): scalable, conversational, compliant with global standards, and built to protect trust.

Final thought

The 2025 Code isn’t just regulation-it’s a reminder that trust is the currency of research. AI can help us earn more of it, but only if we commit to ethics as rigorously as we commit to insights. Researchers who embrace these standards won’t just avoid risk-they’ll future-proof their practice, differentiate their offering, and, most importantly, keep the public on our side.

Because without trust, there is no data, and without data, there are no insights.

‍