Submission on the review of model defamation provisions | Stage 2 Part A

By Fabienne Sharbanee

21 October 2022

The process of reforming defamation law in Australia commenced around 2 years ago.  The review of the Model Defamation Provisions is being managed by the Victorian and New South Wales State governments, with submissions and input sought from stakeholders around the country.

Stage 2 of the reform process began in September 2022, and was divided into Parts A and B.  Part A addressed the liability of internet intermediaries for third-party content.  

Bennett made its submission to the NSW working party on Part A of Stage 2 of the reform process in September 2022. 

You can read the full content of the submission below, or you can access it along with all other public submissions here.

---

Bennett submission on Part A of Stage 2 of the defamation reform process – Liability of internet intermediaries

PART A – Liability of Internet Intermediaries

Note on submissions by others

1. Bennett’s contributing authors have had the benefit of reading the “Submission by Defamation Lawyers” authored by Sue Chrysanthou SC and others (Defamation Lawyers’ Submission)¹, which we support. In particular, we strongly endorse the authors’ proposed amendments to section 32 (provided in Part D at [64]) as an alternative to draft sections 9A, 31A and 39A.
2. That proposed amendment simply and efficiently balances the rights of plaintiffs and internet intermediaries, while ensuring that international megacorporations are granted the same protections as everyone else.
3. What follows below are Bennett’s general comments and high-level submissions in respect of each recommendation addressed in the Background Paper²

General comments

Google LLC v Defteros

4. The Background Paper to these proposals explains that the broad test for publication is the premise of the proposed recommendations. Martin Bennett and Michael Douglas previously wrote on that broad test in: ‘“Publication” of Defamation in the Digital Era’ (2020) 47(7) Brief 6.

5. That test was changed by the High Court in Google v Defteros [2022] HCA 27 (Defteros). Internet intermediaries are no longer prima facie publishers of defamatory material.

6. The recommendations are underpinned by the proposition that the current law strikes the wrong balance between protecting reputation on the one hand, and enabling freedom of expression and access to information on the other. Following Defteros, the balance has changed. It does not need to be shifted further to the benefit of intermediaries.

7. The High Court’s decision in Defteros undermines the logic of the entirety of Part A of Stage 2 of the Review.

“Agreement” among stakeholders

8. There may be agreement among stakeholders whose business models damage reputation and who face the prospect of defamation liability to the effect that the proposed recommendations are justified.

9. There can be no such agreement among stakeholders on the other side of the issue. The problem is that “plaintiff” stakeholders are contingent—they only become stakeholders once they suffer reputational harm. This is a point we stressed in our previous submission to this review.

10. The democratic premise underlying the policy of the recommendations is seriously flawed in circumstances where the views of “defendant” stakeholders are given disproportionate weight. Ignoring or downplaying the views of plaintiff-aligned stakeholders in enacting the Stage 1 reforms has to date only increased the costs and complexity of defamation disputes to the detriment of ordinary Australians.

Recommendation 1: Conditional, statutory exemption from defamation liability for mere conduits, caching and storage services

11. While at first blush there appears to be a sound policy basis to statutorily exclude internet service providers from liability, Recommendation 1 is superfluous.

12. First, internet intermediaries including ISPs and caching and storage services are generally not the subject of defamation claims.

13. Secondly, mere conduits, caching and storage services would not meet the test of “publication” following Defteros.

14. Thirdly, s 235 of the Online Safety Act 2021 (Cth) already provides immunity from liability in defamation to internet service providers and Australian hosting service providers “where the provider was not aware of the nature of the online content”³. This is sufficient, and further piecemeal statutory exemptions are unnecessary.

Recommendation 2: Conditional, statutory exemption from defamation liability for standard search engine functions

15. Following the High Court’s decision in Defteros, this recommendation is also superfluous. Search engines are no longer prima facie considered publishers of the defamatory online content to which they provide hyperlinks.

16. Notably however, Keane and Gordon JJ in dissent in Defteros recognised that a person who aids another to comprehend defamatory matter does participate in the publication of that matter to that person ⁴. A search engine such as Google facilitates publications that may be defamatory, by presenting search results to afford access to webpages ⁵. Their Honours considered that search engines cannot be accurately described as a passive instrument by means of which primary publishers convey information⁶, as they have an active interest in the content published.

17. Search engines, particularly Google, provide functionality beyond the neutral provision of search results⁷, and do have control over content. Although no longer considered “publishers” of organic search results, they undoubtedly play an active role in the dissemination and amplification of material, which impacts the extent of reputational damage.

18. In our experience, the removal of links to defamatory publications is a remedy highly sought after by plaintiffs, and is necessary in order to properly vindicate their reputation. Bennett supports the introduction of a “right to erasure” similar to art 17 of the GDPR⁸.

Recommendation 3A: Model A – Safe harbour defence for digital intermediaries, subject to a simple complaints notice process (Alternative to Recommendation 3B)

19. It is highly unlikely that an originator would ever consent to their contact details being shared with a complainant. Our experience is that even journalists employed by major media outlets do not cooperate when their service details are sought by plaintiffs.

20. The draft Model Amendments effectively provide that an originator can only consent to the intermediary providing their contact details after a complaint is made⁹. There is no scope for ongoing or advanced consent to be required as part of an intermediary’s terms and conditions of use.

21. As we noted in our previous submission¹⁰, in our experience, the vast majority of complainants who seek to engage with internet intermediaries in respect of third-party content do so not to seek compensation or redress, but because they want to prevent further harm to their reputation from the accessibility of defamatory material on the internet.

22. Model A effectively requires extra work on the part of an intermediary (to contact and seek the consent of an originator) just to arrive at the same result as Model B: the removal of access to defamatory publications after a complaint is made. However, it is worse in that it removes all incentive to remove or block access to defamatory material in circumstances where the original poster can be served.

23. For these reasons, Bennett supports Model B over Model A. However, as noted above, the preferable course is to extend the defence of innocent dissemination in the manner proposed in the Defamation Lawyers’ Submission, or alternatively (as described below), to a hybrid of the two.

Recommendation 3B: Model B – innocent dissemination defence for digital intermediaries, subject to a simple complaints notice process (Alternative to Recommendation 3A)

24. Model B is the preferable option of the two proposed as it provides a relatively fast and simple method to obtain a remedy for online defamatory content – something which has long been sought after in this space.

25. There is no need for a “safe harbour” when digital intermediaries can instead avail themselves of an innocent dissemination defence.

26. In our view, the 14 days allowed by draft section 31A(1)(c) is too long; 7 days (at most) would suffice. Experience and history demonstrate that people who are defamed by online publications suffer the greatest reputational damage in the period immediately following publication. There should be no reason why a digital intermediary cannot take access prevention steps within 48 hours of receipt of a complaints notice.

27. The Background Paper notes that the risks of Model B include:

27.1 It does not provide the poster with the opportunity to defend the content; and

27.2 It may create an incentive for a complainant to approach an internet intermediary to have legitimate content removed, rather than approach the poster who may refuse or defend the content. This could result in over-censorship and the removal of legitimate content.

28. The ideal solution is to instead adopt the amendments to section 32 proposed by the Defamation Lawyers’ Submission rather than implement Model A or Model B. However, if it is considered that some form of complaints notice process is necessary, an alternative hybrid model would overcome these concerns.

29. For example:

29.1 Upon receipt of a complaints notice, an internet intermediary would be required to immediately take down or block access to the (allegedly) defamatory publication;

29.2 The internet intermediary would then be required to seek the original poster’s consent to provide their contact details to the complainant;

29.3 Access to the publication could be reinstated if the original poster’s contact details are provided to the internet intermediary and passed on to the complainant within 12 months of the date of first publication.

30. This would not affect legitimate news media as such publishers are readily identifiable; but it limits damage to a complainant’s reputation, provides a mechanism to connect the complainant with an anonymous original poster, and the original poster is afforded an opportunity to defend the content of their publication.

Recommendation 4: Commonwealth Government to consider an exemption for defamation law from the Online Safety Act 2021 immunity

31. Recommendation 4 is entirely unnecessary. The Background Paper itself summarises the position as follows:

“…an exemption from the OSA provisions is not strictly necessary, but nevertheless, it may be desirable to provide clarity to litigants. Such an exemption would make it very clear that defamation law does not require reference to the OSA, and potentially avoid complex disputes in litigation which test the issue.”¹¹

32. This is not a strong reason for an exemption from the OSA. The application of the OSA is quite clear and introducing an exemption risks introducing ambiguity.

33. The Background Paper also notes that “if Recommendation 1 is introduced, this would make the application of the OSA immunity to these internet intermediary functions irrelevant in the context of defamation law.” ¹²

34. It appears from this that an exemption is required in order to give Recommendation 1 work to do. This would add additional complexity to the analysis of a client’s case, increasing costs and therefore diminishing the likelihood of a speedy, non-judicial settlement of a dispute, contrary to the purposes of the Uniform Defamation Acts. This is a further reason against implementing Recommendation 1.

35. Legislative provisions which deal with the same subject matter in different statutes rarely (if ever) result in greater clarity for litigants, but rather, in greater ambiguity, and scope for disputation.

Recommendation 5: Empower courts to make non-party orders to prevent access to defamatory matter online

36. Proposed section 39A (Recommendation 5) is a useful addition to the Uniform Defamation Acts in that it provides a clear statutory basis for an access prevention order, particularly in the event that Model A is adopted.

37. However, by providing a “safe harbour” as envisaged by Model A, there is no incentive for internet intermediaries to voluntarily co-operate or act quickly to mitigate damage to a plaintiff’s reputation, particularly if the ‘original poster’ is easily identifiable.

38. As we noted in our previous submission¹³, the real issue is an Australian court’s jurisdiction to make such orders and the cost of enforcement. If internet intermediaries are to benefit from any new defence under these amendments, it should only be in circumstances where they must submit to the jurisdiction of Australian courts as the price of deriving income by providing services in Australia.

Recommendation 6: Courts to consider balancing factors when making preliminary discovery orders

39. At the outset it is clear that draft s 23A would not apply in the Federal Court¹⁴. Rule 7.22 of the Federal Court Rules lists the matters the Court must be satisfied of before ordering preliminary discovery. Enacting this amendment would probably increase the likelihood of “forum shopping”, not prevent it.

40. Further, courts are already able to – and do – balance the relevant factors referred to in draft s 23A. For example, see NW v Bechtel (Western Australia) Pty Ltd [2014] WASC 375 in which Master Sanderson refused an application for pre-action discovery on such grounds despite the applicant otherwise meeting the relevant criteria under the Supreme Court Rules 1971 (WA).

Recommendation 7: Mandatory requirements for an offer to make amends to be updated for online publications

41 The Background Paper notes that the draft amendments to s 15 are intended to come into play where an internet intermediary has received a complaints notice and has chosen to leave the defamatory content online, thus losing the benefit of the defence provided under either Model A or Model B¹⁵.

42 If an internet intermediary is going to stand by an anonymous defamatory publication to the point that a complainant is put to the trouble (and consequent delay and cost) of issuing a complaints notice and a concerns notice, any offer to make amends ought to include an offer to take the steps referred to in s 15(1)(d) and (e).

43. However, the draft amendment is not limited to internet intermediaries. It instead refers to “digital matter”. If such an amendment is to be adopted, it ought to read (for example):

s 15(1B): If the matter in question is a digital matter If the publisher in question is an internet intermediary:

(a) an offer to make amends may, instead of making the offer mentioned in subsection (1)(d), include an offer to take access prevention steps, and

(b) the offer mentioned in subsection (1)(e) is not required to be made if an offer to take access prevention steps is made.

44. A publisher of digital matter who is not a mere internet intermediary ought to still offer to take access prevention steps (if possible) as provided by draft s15(1A)(b) but should not receive the benefit of draft s 15(1B)

Media Law Group

Contributing authors: Martin Bennett, Fabienne Sharbanee, Taleesha Elder and Michael Douglas

1. 3 of the contributing authors of this submission (Martin Bennett, Fabienne Sharbanee and Michael Douglas) are also signatories to the Defamation Lawyers’ Submission.
2. State of New South Wales, Background Paper: Model Defamation Amendment Provisions 2022 (Consultation Draft), August 2022 (Background Paper)
3. See also Fairfax Media Publications v Voller (2020) 105 NSWLR 83 in relation to this protection as previously afforded by Schedule 5 cl 91 of the Broadcasting Services Act 1992 (Cth)
4. Defteros, [95]
5. Defteros, [98]
6. Defteros, [100]
7. Such as ‘snippets’ or autocompletion of suggested searches
8. Regulation (EU) 2016/679 (General Data Protection Regulation)
9. Draft subsection 31A(2) [Part A, recommendation 3A alternative]
10. Bennett + Co, Response to Discussion Paper, Attorneys- General Review of Model Defamation Provisions – Stage 2, 19.05.2021
11. Background Paper, 46
12. Background Paper, 45.
13. Bennett + Co, Response to Discussion Paper, Attorneys-General Review of Model Defamation Provisions – Stage 2, 19 May 2021 at [15(b)]
14. See s 79(1) of the Judiciary Act 1903 (Cth) and r 7.22 of the Federal Court Rules 2011 (Cth)]
15. Background Paper, 57

Disclaimer: The information published in this article is of a general nature and should not be construed as legal advice. Whilst we aim to provide timely, relevant and accurate information, the law may change and circumstances may differ. You should not therefore act in reliance on it without first obtaining specific legal advice.