Report
Lighthouse Reports says it has taken a demand for environmental data to the Aarhus compliance process. The dispute turns on the difference between efficiency ratings and a community’s ability to see actual resource use.
Lighthouse Reports published a cross-border investigation on September 30 and said it had submitted a communication to the Aarhus Convention Compliance Committee seeking disclosure of data-center environmental information. The newsroom describes a year of information requests to European authorities and challenges the extent to which the public can examine facility-level energy and water figures. Its filing is a claim for review, not a ruling that the European Commission has broken the law. Source: lighthousereports.com
In its accompanying explanation, Lighthouse argues that the Commission’s new efficiency labels leave out consumption figures communities need to assess local impacts. AI Faith Monitor has read that published explanation but has not independently audited the correspondence or obtained a committee decision on the case. Source: lighthousereports.com
The dispute comes as Europe tries to expand computing capacity while managing its resource demands. It raises a practical question for residents, public purchasers and civil society: what must a disclosure system reveal before people can evaluate the infrastructure being proposed around them?
The Commission’s September 21 delegated-act text explains a different emphasis. It presents standardized electronic labels as a way to compare data-center performance and encourage more efficient designs. The scheme draws on information operators already report, with the stated aim of improving comparability without adding a new reporting burden. Its explanatory memorandum explicitly invokes protection of trade and business secrets. Source: energy.ec.europa.eu
That is a substantive argument for a common rating system. Consistent categories can make complex performance data easier to interpret, including for customers choosing among computing providers. They can also help distinguish a facility’s efficiency from the sheer size of its operation.
But a comparison tool and a local resource assessment answer different questions. An efficient facility can still impose a large total demand. Conversely, a smaller total demand does not establish that a facility uses each unit of energy well. Readers need to know which question a label is designed to answer before treating it as a complete environmental verdict.
Consider a hypothetical pair of facilities. One requires less supporting energy for every unit used by its computing equipment, yet operates at a much larger scale. The other has a less favorable efficiency ratio but uses less electricity overall. A ranking based on the ratio would help compare operating efficiency. It would not tell a neighboring town how much additional electricity the larger facility needs.
Water raises another distinction. An annual total can help establish scale, but its local meaning also depends on the source and timing of withdrawals. A figure that is small in a national account may matter differently during a dry period in a specific supply area. These examples explain the information problem; they do not establish damage at any named European site.
Publication of useful figures also needs clear units, reporting periods and boundaries. Comparing figures with different definitions can create an appearance of accountability while leaving the real question unanswered.
Statistics Netherlands provides an example of what a clearly bounded public statistic can establish. In December 2025, it reported that data centers used 5,100 gigawatt-hours of electricity in 2024, around 4.6 percent of Dutch electricity consumption. Its release identifies the 2024 figures as provisional and limits the dataset to electricity connections where a data center is the main activity. It excludes facilities embedded in institutions such as universities and hospitals. Source: cbs.nl
Those qualifications are essential. The figure does not measure AI electricity consumption separately, and it does not describe a particular neighborhood’s grid conditions. It does establish that computing infrastructure is large enough to warrant serious public accounting.
The release also shows why a national total cannot replace site-level information. Knowing the aggregate share helps frame the debate. Evaluating a proposed connection, cooling system or local planning decision requires a different level of detail. Neither dataset makes the other redundant.
UNECE’s implementation guide describes access to information, participation in decision-making and access to justice as the Aarhus Convention’s connected pillars. It distinguishes information supplied on request from authorities’ duty to disseminate information proactively. The guide explains the framework; it does not decide this new complaint. Source: unece.org
UNECE’s explanation of the convention also recognizes that commercial confidentiality can be a legitimate interest while requiring restrictive interpretation and attention to the public interest in disclosure. Applying those principles to particular data, exceptions and decisions is the work the compliance process would have to do. This report does not pre-empt it. Source: unece.org
There is a separate near-term policy opportunity. The Commission opened consultation on possible minimum data-center performance standards on September 21, with responses invited until December 14. Its announcement says a legislative proposal is planned for the second quarter of 2027. Those standards are still being developed. Source: energy.ec.europa.eu
In 2 Corinthians 8:16–24, Paul explains practical arrangements for administering a collection, including trusted companions and concern for how its handling is seen by others. The setting is church generosity, rather than environmental regulation. The relevant principle is that a good purpose does not remove the value of visible, accountable administration. 2 Corinthians 8:16-24 (NIV)
For Christians concerned about the AI build-out, that supports asking for intelligible evidence without presuming bad faith. A church purchasing digital services can ask what environmental claims cover and which figures support them. A community group can seek information that helps residents participate meaningfully, while accepting that some genuinely sensitive details may need protection.
The current dispute should be judged through its documents and eventual findings. Its broader challenge is already clear: an assurance of responsible stewardship becomes more useful when affected people can examine the evidence behind it.
Lighthouse Reports: Data Centre Silence
Reports the newsroom’s published account of its complaint, not an independently confirmed docket entry or a finding of illegality. Commission delegated-act text and annex inspected. The original correspondence and complete underlying operator dataset were not independently audited. Dutch electricity figures cover 2024 and are not an AI-only total.
News brief / Policy & power
Today’s announcement raises a practical question for religious institutions using AI: what would make an assurance independently checkable?
Published 2026-09-29 · Source 2026-09-29
News brief / Policy & power
Remarks published September 24 connect access to AI with dignity, public responsibility and meaningful participation.
Published 2026-09-28 · Source 2026-09-24
Report / Policy & power
The ecumenical gathering points toward a role for churches beyond adopting tools for their own offices.
Source / event 2026-04-14 · Updated 2026-09-27