Why removal will become rarer

The right to be forgotten, as a practical matter rather than a legal concept, is contracting. A decade ago, a careful approach to a search engine, a careful approach to an article's publisher, and a careful approach to a few caching services could substantially reduce what an ordinary search returned about a person. Today, the same effort produces a more partial result.

The reasons are several. The number of sources has grown. A removal from one search engine is no longer a removal from the searchable record; alternative search engines and aggregators, including those built on different infrastructure, may continue to return the content. The original publisher's removal is no longer a removal from the wider record; archival mirrors, opportunistic re-publishers, and AI training corpora may have already preserved it.

The European framework that gave the right its legal form is itself under pressure. Subsequent rulings have refined the balance between privacy and freedom of information in ways that, in practice, narrow the cases in which a removal is granted. The standard of proof for a serious removal request is higher than it was, and the categories of information that qualify have narrowed.

The economic incentive of those holding the records has shifted. AI training has created a market in archived content. Records that were previously held but unused are now valuable as training data, and the platforms that hold them are less inclined to delete than to retain. A removal request now reaches an institution that has, in the meantime, monetised the underlying record.

The desk treats removal as one tool among several, not as the primary tool. The work of keeping a settled position is largely the work of managing what is around the residual record rather than removing the record itself.