Related items

The European Commission’s forthcoming guidance on the Corporate Sustainability Due Diligence Directive (CS3D) should:

• allow companies to voluntarily prioritise chains of activities with a meaningful connection, which should be further defined in the guidance, to the EU or based on a global risk-based approach in cases where companies have global integrated supply chains;

• recognise that companies may face conflicting or overlapping legal requirements between EU and third-country legislation, particularly in relation to information gathering, audits, data transfers, supplier disengagement and cooperation with authorities;

• clarify how companies should document and manage circumstances in which third-country law restricts or prevents a due diligence measure;

• recognise interactions with competent local authorities, regulatory inspections, permits, licences and other official approvals as potentially relevant sources of due diligence information; and

• protect companies from liability where they have followed a reasonable, documented and good-faith process, including where another stakeholder might have prioritised risks or selected due diligence measures differently.

• If these measures are included in the guidance, they would help the CS3D deliver meaningful and effective due diligence, rather than an exhaustive mapping of every global business relationship. Flexibility, proportionality and legal certainty can help companies progressively develop credible global systems while directing resources towards the most significant risks and the areas where they have the greatest ability to achieve positive outcomes.

Learn more in our contribution and paper.

Meta’s open-source Canopy Height Maps helped the European Commission’s Joint Research Centre improve its 2020 Global Forest Cover map, demonstrating how artificial intelligence can strengthen Europe’s ability to monitor forests and advance climate and biodiversity goals. Developed with the World Resources Institute, the maps use Meta’s DINO self-supervised computer vision model to estimate tree height from satellite imagery. This helped researchers apply the five-metre threshold used in many international forest definitions and better distinguish forests from other tree-covered land, including agricultural tree crops. The resulting product, available through the EU Forest Observatory, provides one of the most comprehensive open views of global forest cover. Read more on Invested in Europe.