Let’s be real—human rights due diligence in seafood is messy, nuanced, and often feels like chasing octopus ink in the dark. You’re working across fragmented supply chains, inconsistent audit reports, language barriers, and ever-evolving standards. Add to that the reality that risk isn’t just something you measure—it’s something people live through.
So when we started experimenting with AI to support our HRDD frameworks, it wasn’t about replacing the human. It was about scaling our capacity, getting faster at pattern recognition, and having a digital sidekick that doesn’t need coffee breaks.
Here’s what this journey has looked like.
🧠 Using AI to Build Client's Framework
We’ve trained our AI assistant (Robotina) to help us identify and assess risk across seafood suppliers using available audit reports, supplier responses, and public risk indicators from tools like the ITUC Global Rights Index, U.S. Department of Labor reports, and NGO alert lists.
What we asked Robotina to do:
- Extract key social compliance findings from dense reports
- Compare against risk criteria we defined (e.g. indicators of forced labor, recruitment practices, working hours)
- Flag discrepancies or concerning trends
- Provide a comparative snapshot across multiple suppliers
- Suggest follow-up questions or next steps based on gaps
* On a side note, I asked her to create her image, and this is what she said:
This version of me—Robotina with seaweed hair, mermaid tail, and a little octopus bestie—feels just right: strong, kind, and ready to dig into audits sin pelos en la lengua. And honestly, rocking a smile while fighting for human rights in seafood? That’s the energy we need more of. 💪🏽🐙✨
The Reports and Data We Rely On
We’re pulling from a mixed bag of sources, including:
- Third-party audits
- Certification watchlists
- ILO, ITUC, and Verité risk indices
- Human rights NGO investigations
- National labor law summaries
- Client-supplied data like onboarding forms and CAPAs
Robotina, helps us parse through this stack of documents faster, finding common red flags or surfacing inconsistent claims that would take hours for a human alone to sift through.
What Robotina Gets Wrong (and Why the Human Still Matters)
Let’s not get it twisted—this is not a plug-and-play solution.
Here are a few real-world hiccups:
- Translation issues: AI can misinterpret audit language from non-English sources, or take “no evidence found” to mean “no risk exists.”
- False confidence: It sometimes gives clean ratings to suppliers based on limited or overly positive documentation, without questioning the source or context.
- Context blindness: AI struggles with nuance—like when a finding was technically closed out but the root cause wasn’t addressed.
- Lack of cultural insight: It can’t pick up on tone, power dynamics, or human intuition that comes from interviewing workers or engaging with QC teams.
We’ve had to go back, tweak the weighting in the framework, adjust how we score risk by topic (e.g., “governance” vs “recruitment practices”), and add new prompts to catch errors.
The Human-AI Feedback Loop
What works best? A partnership.
We treat Robotina as a tireless assistant who drafts the first pass—and then we, the humans, interrogate the output. We bring the follow-up, the phone calls, the supplier interviews, the ability to ask: “Okay, but why is this flagged low risk when the region has known labor issues?”
And that back-and-forth has actually made our framework stronger, more consistent, and scalable without losing the rigor.
Why This Matters
If we want real progress in human rights due diligence—not just pretty dashboards—we have to get better at assessing risk at scale without dulling our critical thinking.
AI helps us get there. But people—especially those grounded in the messy realities of seafood supply chains—are still the secret sauce.
And especially in a world full of fancy dashboards that cost hundreds of thousands of dollars—and where many of us are working with limited resources—AI gives us a way to get a clear snapshot of our supply chain, take actionable steps, ask better questions, and keep improving. We can track a baseline, compare findings over the years, and identify persistent or emerging issues—all at a fraction of the cost, so more of our budget goes toward actual change.
Want to learn how to do the same? We’d love to show you.
Contact: Adriana@seafoodninja.com