In the Board’s view, the principle of free evaluation of evidence applies, thus leading to a decision on the balance of probabilities. The Board holds that there is no legal basis for a stricter standard of proof in respect of prior art originating from the internet.

Internet publications play an ever increasing role in determining the relevant prior art and thus cannot be ignored. However, unlike conventional paper publications, internet disclosures present specific difficulties, notably in that they may have been modified in the course of time without such modifications being reliably traceable. These immanent difficulties induced Board 3.2.04 to require a strict standard of proof (“beyond any reasonable doubt” or “up to the hilt”) for public availability of internet disclosures in its decision T 1134/06 “Internet citations/KONAMI” of January 16, 2007. T 1134/06 has been affirmed by Board 3.3.04 in decisions T 1875/06 and T 1213/05 and by Board 3.3.05 in T 19/05, whereas decision T 2339/09 by Board 3.2.03, though referring to T 1134/06, relied on a standard of “balance of probabilities”. Furthermore, the “up to the hilt” standard was explicitly rejected by Board 3.3.08 as not applicable to internet disclosures in its decision T 990/09. Hence, despite the issue’s topicality, there is a sustained uncertainty arising from the existing case law in respect of proof requirements for internet disclosures.

In the recent case T 286/10 (in French), Board 3.5.06 took this issue up again and declared any specific proof requirements for internet disclosures to have no legal basis. In the Board’s view, in order to remedy the difficulties associated with internet disclosures, it was not necessary to derogate from the application of the common standard of proof according to which the existence of a prior publication is to be established by evaluating the probabilities according to the circumstances of the case. In this respect, the Board follows decisions T 2339/09 and T 990/09 under the premise of free evaluation of evidence as explained in the earlier decision T 750/94.

In order to justify a strict standard of proof for internet disclosures, T 1134/06 referred to decision T 472/92. However, T 472/92 also confirmed the constant jurisprudence of the boards of appeal, according to which the evidence was generally evaluated in accordance with what seemed most probable, and only made an exception when a prior use was alleged where practically all the evidence lay within the power and knowledge of the opponent. In the present case, the Board did not consider the circumstances of internet disclosures to be comparable to a prior use situation so that T 472/92 could not be relied on to strengthen the proof requirements.

Moreover, T 1134/06 itself distinguished cases of internet publications where in fact a decision on the balance of probabilities was to be taken, e.g. publications originating from internet sites belonging to a reliable editor of good repute who published online versions corresponding to paper publications. In such cases, the content and date of internet disclosures did not need any supplementary proof.

All in all, the Board concluded that, as a matter of principle, internet publications did not require a different standard of proof; the uncertainties associated with these disclosures had to be removed by providing a sufficient degree of probability and establishing a presumption of public availability to the conviction of the judge. There was no reason to increase the degree of probability to the extent of lack of any reasonable doubt.

In the present case, of the two internet publications A9 and A10 advanced by the appellant, article A9 was retrievable at the same internet address to this day andprima facie resulted from a known and reliable source of information. Hence, its date of publication seemed highly probable.

From the URL indicated on A10, it had to be inferred that this document had been archived by the Internet Archive on a date before the priority date of the patent-in-suit. The Board did not arrive at the conclusion of T 1134/06 that internet archives were basically not reliable. In particular, the Board held that normally the mere fact that a document had been archived at at a certain date, without any particular circumstance arousing suspicion, already justified the assumption that public access to the document was possible on the date of uploading it to the internet and then via the Internet Archive shortly afterwards.

The Internet Archive, a private non-profit archiving initiative, placed former internet snapshots at the disposal of the public at large. Since its creation in 1996, it has become very popular and acquired a good reputation. Even though the data volume was enormous, the Archive, of course, only comprised an incomplete collection of those former internet pages, which was, however, also the case for classical libraries and did not affect the credibility of the available information. Although the Board did not exclude the possibility of doubts as to individual entries in the Archive, it believed that the Archive itself sufficiently guaranteed the presumption of a reliable and trustworthy source of information so that an opposing party, depending on the case, had to produce evidence casting doubt on that assumed reliability and thus refuting that presumption. Consequently, a general allegation of lack of reliability of the Internet Archive was not sufficient to question the date of public access to an archived document.

As the respondent had not given any explanations as to why the presumption of public availability was not valid, the Board considered both documents to form part of the prior art and revoked the patent for lack of inventive step with respect to this prior art.

Since the importance of internet disclosures for determining the relevant prior art cannot be overestimated, a standard of proof practically excluding this valuable source of information does not appear expedient. On the other hand, the premature and light-minded acceptance of publication dates in a medium of more or less transient nature would interfere with the principles of the patent system where the requirements of novelty and inventive step are to be judged on the basis of the established prior art. T 286/10 attempts to strike a balance between these opposing demands by limiting the “up to the hilt” proof requirement of T 472/92 to the specific prior use case and relying instead on the principle of free evaluation of evidence and a standard of balance of probabilities, i.e., a principle and a standard consistently applied by the boards of appeal in other circumstances. This approach seems to be reasonable: if all the information is exceptionally in the hands of one party, that party may be seen to bear the full onus of proof. If, however, evidence from elsewhere is to be evaluated, all parties are in an equal position to comment on its persuasiveness so as to convince the board.

Determining the dates and content of internet disclosures may indeed present specific difficulties which have to be carefully taken into account. These difficulties, however, are not unique and may also arise with other sources of disclosure, e.g. oral disclosures or standard preparatory documents. In all of these cases, a decision on the balance of probabilities is normally taken.

Whether or not T 286/10 will close the debate on standards of proof for internet publications remains to be seen. In any case, the decision confirms the approach envisaged by the new version of the EPO Guidelines for Examination of November 1st, 2014 (cf. G- IV, 7.5.2).