MCP data hoarding leads to parallel data households, exactly the type of thing we spent a lot of energy on to reduce

n:: Principle of least privilege: applications only should have access to data they need, and never more. Data hoarding in MCPs goes beyond that.

MCP is meant to abstract the way access is created to resources. In practice it gets used to abstract away any decision on which data to provide or not. That's the trap.

https://web.archive.org/web/20241127055007/https://openknowledge.worldbank.org/entities/publication/8a6cb591-4548-5786-b108-5f9cd32edc8c

2021 paper looking at data governance legal frameworks globally in 80 countries, EU and USA are absent (Estonia, UK included though), and lumps Europe and Central Asia together, which leads to phrases like 'UK and Estonia have this, but elsewhere in the region Kyrgyzstan hasn't' so the region is mediocre at best. Apples/pears. Mentions GDPR more or less as the single EU framework here, despite the 2018 free flow of non-personal data regulation which became applicable in May 2019. (others within the EU Data Strategy / single market for data had been announced or proposed by 2021 but not in place and aren't mentioned here either.)

Companies, half of an anecdotal sample of some 20 US CDOs, have turned Copilot off / restricting it strongly. This as it surfaces info in summaries etc that employees would not have direct access to. No access security connection between Copilot and results. So data governance is blocking its roll-out.

Orgalim is an industry association for tech manufacturers, and has been selected as a member of the EDIB working group. Their topics of interest, and thus perspective on governance and standards, is DT for industrial products/manufacturing, digital product passports (relevant to GDDS and in PLM), as well as smaller manufacturing dataspaces (I should come up with a term for non pan-EU generic DSs. Xa Xb Xc etc) Note the mention, and link, of 'net-zero' policy, a warning flag for greenwashing.