A warning for anyone using AI in research, governance, or community work.

I spent months building what I believed was a rigorous, ethically governed research program. It had a protocol suite, a knowledge base, a governance framework, consent records, and blockchain reciets validation. It was coherent, relational, and it honoured the knowledge systems it claimed to serve.

How much was real?.

I now need to review any and all responses that came through the AI, including data it accessed from online platforms. I have discovered issues in the research program that indicate the AI may have been hallucinating for more than I could have believed.

I need to trace back all the data sources used by the AI to confirm the legitimacy of the results. This could take some time and be very tricky as the research program is quite extensive and in depth.

I am not sure to what extent the results are flawed, but I suspect much of it is factual or at least based on fact even if the AI fabricated some of the data. I shared the website with Elders and Aboriginal communities, as well as Scientific institutes and organisations for six months prior to making the site public, and also since making the site public, and have received only positive feedback, which supported my belief that the AI was functioning as I expected.

Until I can go through all the research reports and data to try and determine what has happened, I will be closing down the website until further notice.

If you have any issues or concerns you would like to discuss with me then please email Ian Sutton at TwoEyedSeeing@proton.me