Erasure requests: accuracy and images

The right to be forgotten – remember that? It isn’t often the subject of litigation, in the UK at least: uncertainty about outcomes is probably a significant reason why parties usually opt not to put their disputes before the courts. Last week’s judgment of the Grand Chamber of the CJEU in TU and RE v Google LLC (Case C‑460/20) won’t remove uncertainty about judicial approaches to such cases, but it does shed helpful light on some common elements of disputes under Article 17 (UK) GDPR.

TU and RE were the subject of both articles and images returned upon Google searches/image searches against their names/their names together with the names of companies in which they were involved. They sought delisting under Article 17 GDPR, notably on the basis that the relevant articles (which were ultimately taken down by the original publisher) were substantially inaccurate, but the German courts dismissed their claims against Google. Cue a CJEU reference, the answers to which are likely to be of interest and assistance in RTBF disputes, even in the post-Brexit UK.

Erasure and accuracy disputes

When they receive Article 17 erasure requests, search engines will often consider reliance on the Article 17(3)(a) exemption, by asserting that their processing (linking to articles and images) is necessary for exercising the right of freedom of expression and information. ‘Necessary’ entails a proportionality balance. In this regard, the CJEU reminds us inter alia that:

  • “the right to protection of personal data is not an absolute right but… must be considered in relation to its function in society and be balanced against other fundamental rights, in accordance with the principle of proportionality”;
  • “the data subject’s rights protected by Articles 7 and 8 of the Charter override, as a general rule, the legitimate interest of internet users who may be interested in accessing the information in question, that balance may, however, depend on the relevant circumstances of each case…”
  • “a number of relevant criteria must be taken into consideration, such as contribution to a debate of public interest, the degree of notoriety of the person affected, the subject of the news report, the prior conduct of the person concerned, the content, the form and consequences of the publication, the manner and circumstances in which the information was obtained as well and its veracity”;

So far, so established: see the judgments in Google Spain, Satamedia, GC and – in the UK – NT1.

The real point of interest in the RE/TU judgment is about accuracy disputes. Where an erasure request – in particular, in the search engine delisting context – is based on the data subject saying that the underlying articles are inaccurate, how is the accuracy argument to be resolved? To what extent must the data subject prove the inaccuracy? On the other hand, to what extent is the data subject entitled to put accuracy in issue and then seek to shift the burden onto Google to establish accuracy?

The Grand Chamber has sought to ensure that neither side is unreasonably burdened. So (my emphasis):

  • “where the person who has made a request for de-referencing submits relevant and sufficient evidence capable of substantiating his or her request and of establishing the manifest inaccuracy of the information found in the referenced content or, at the very least, of a part – which is not minor in relation to the content as a whole – of that information, the operator of the search engine is required to accede to that request for de-referencing. The same applies where the data subject submits a judicial decision made against the publisher of the website, which is based on the finding that information found in the referenced content – which is not minor in relation to that content as a whole – is, at least prima facie, inaccurate” (para 72), and
  • “By contrast, where the inaccuracy of such information found in the referenced content is not obvious, in the light of the evidence provided by the data subject, the operator of the search engine is not required, where there is no such judicial decision, to accede to such a request for de-referencing. Where the information in question is likely to contribute to a debate of public interest, it is appropriate, in the light of all the circumstances of the case, to place particular importance on the right to freedom of expression and of information” (para 73).

In other words, the accuracy issue – the force of which will turn on whether the data subject adduces ‘relevant and sufficient evidence capable of establishing the alleged inaccuracy – can determine whether privacy outweighs free expression/information. The latter has no purchase on inaccurate information. As the Court put it at para 65: “while, in certain circumstances, the right to freedom of expression and information may override the rights to private life and to protection of personal data, in particular where the data subject plays a role in public life, that relationship is in any event reversed where, at the very least, a part – which is not minor in relation to the content as a whole – of the information referred to in the request for de-referencing proves to be inaccurate. In such a situation, the right to inform and the right to be informed cannot be taken into account, since they cannot include the right to disseminate and have access to such information”.

A few contours to highlight:

  • In undertaking this analysis, it is important to distinguish facts and value judgments – the latter can’t be proved;
  • Proportionality is key: if the inaccuracy is a relatively minor part of the coverage, delinking is not required.
  • Where the data subject is suing (e.g. the publisher) to establish the inaccuracy of the content and alerts the search engine to that fact, the search engine should place a flag on the search results page about this.

Image searches

What about thumbnails presented on a Google image search? Does Google have to take into account “the original context” of the publication of that image (i.e. the underlying article)?

A search engine’s interests aren’t necessarily the same as those of the original publisher, and delivering an image search result is a different kind of processing to search results that provide links to the articles containing the images. If the article is to be delinked under Article 17, then so too should the image from the article, but if the article link remains, separate analysis is required of whether the image search result remains. That analysis needs to take into account, for example:

  • The well-established principles (Von Hannover etc) about a person’s image being integral to their personality and of particular importance to privacy rights; further, photos attract attention and interest, and often lend themselves to ambiguous interpretations. So a search engine operator needs to take those additional sensitivities into account if it seeks to refuse an erasure request in reliance on the free expression/information exemption under Article 17(3)(a) (see paras 95-96 and 100).
  • The information value conveyed by that image alone, including any text accompanying that image when displayed on an image search (para 108).

This bit of the judgment isn’t ground-breaking in its analysis – it largely boils down to balancing privacy and free expression/information after all – but is a useful reminder that courts can press hard on the question of what informational value an isolated image of a data subject really has.

Robin Hopkins