Regulation

' When the Digital Services Act Goes Global' by Anupam Chander in  Berkeley Technology Law Journal (Forthcoming) comments 

The European Union’s Digital Services Act (“DSA”) establishes a “meta law”—public regulation of the private regulation conducted by internet platforms. The DSA offers an attempt to balance private technological power with democratic oversight. The DSA will likely prove an attractive model for other governments to assert control over massive global internet platforms. What happens when other countries borrow its approach, in an instantiation of the vaunted Brussels Effect? This Article evaluates the DSA using the “Putin Test”—asking what if an authoritarian leader were given the powers granted by the DSA? The Article argues that authoritarians might well exploit various mechanisms in the DSA to enlarge their control over the dissemination of information, and, in particular, to target the speech of critics. 

Kantorowicz

'The failures of political prophecy: Ernst Kantorowicz’s wartime lectures' by Bennett Nagtegaal in (2023) Intellectual History Review comments 

This paper introduces a series of lectures Ernst Kantorowicz offered to the Army Specialized Training Program in 1943 in order to reconsider the development of his intellectual biography. These “wartime lectures” constitute Kantorowicz’s only sustained discussion of modern German history and his only intellectual engagement with Nazism. Introducing these lectures thus presents an opportunity to re-examine the relationship between Kantorowicz’s early and mature works through his assessment of Nazi Germany. For Kantorowicz, Nazism was the violent result of a German commitment to political prophecy. At the core of Kantorowicz’s lectures was a criticism of political theology and its role in modern history. In making this criticism, Kantorowicz simultaneously distanced himself from the prophetic register of his earlier writings. Moreover, recovering Kantorowicz’s concern with modern political theology is also important in foregrounding the intellectual genealogy of The King’s Two Bodies, a work often separated from its more telling subtitle: A Study in Medieval Political Theology. Together, this paper argues that the most significant changes in Kantorowicz’s writings can be traced to the intellectual circumstances of the Second World War.

Gender and Soft Decertification

'Abolishing legal sex status: The challenge and consequences of gender-related law reform', the 2022 final report of the UK Future of Legal Gender Project 2018-2022 states 

 1. The Future of Legal Gender was a collaborative research project, funded by the Economic and Social Research Council, UK, that ran from May 2018 until April 2022. It explored, from a social justice perspective, the legal, social, and policy implications of reforming the current system in England & Wales which requires everyone to have a legal sex. 

2. Having a legal sex begins with birth registration as female or male and continues over a lifetime unless a person formally transitions. Obtaining a Gender Recognition Certificate under the Gender Recognition Act 2004 changes a person’s gender and their legal sex. Otherwise, the presumption, in law, is that a person’s gender is the sex they are registered with at birth. 

3. Legal sex and gender contribute to who we are as legal subjects. They affect how we are treated, and the opportunities that we have, as this report explores. More generally, legal sex status contributes to the social development of women and men as two separate groups of people. It suggests that both sex and gender matter – not simply for remedying inequality but as core settled aspects of who we are. 

4. To explore whether the current system of assigning people a legal sex and gender status should be dismantled, and the challenges and potential difficulties this proposal raises, we undertook extensive research, involving several different methods (described in more detail in the appendix). This included a survey eliciting over 3,000 responses; 200 interviews with government officials, trade unions, regulatory bodies, community organisations, service providers, academics, lawyers, and general publics; and iterative focus group discussions and workshops with lawyers, academics, legal drafting experts, NGOs, and public officials to explore the principles of decertification emerging from our research. 

5. Our research identified benefits to decertification. These included: dismantling a legal system which formally places people, from birth, in unequal social categories of female and male; supporting greater self-expression – free from gender constraints; and removing the legal burdens currently placed on people who want state recognition of a change in their sex and gender status. 

6. Concerns about decertification also emerged from our research. These concerns mainly related to gender and sex-specific services, data collection, violence, and positive action. Some research participants worried that measures to abolish sex as a legal status would make it harder to retain provision and spaces based on distinctions between women and men (or females and males) and that this would disadvantage women. 

7. Our research identified some ways of tackling these concerns. These strategies build on current practices of ‘soft decertification’ as public bodies and other organisations and agencies respond to users, staff, and clients who self-identify outside of a binary framework of gender anchored in the sex registered at birth. However, the hollowing out of legal sex has also faced opposition from groups who assert the importance of attending to women as a class defined by their sex. During this research, public bodies described how they navigated tensions between these competing demands, amid divergent interpretations of the relevant law. 

8. Advancing gender equality as a broad, intersectional agenda does not just depend on state action. It also does not depend on equality law alone; other laws also shape gender relations and whether people can live in gender nonconforming ways. However, equality law has become a site of intense debate. Our research explored how different categories in equality law operate and questioned whether people need to be legally assigned to a category, such as gender or sex, to access legal remedies. Other equality grounds, such as race and sexual orientation, operate without requiring these ‘protected characteristics’ to be part of a person’s legal identity. 

9. Several interviewees suggested that the present political climate was not a suitable or safe one in which to question the architecture of equality law or to radically alter gender and sex categories. Decertification may therefore be better approached through the prism of ‘slow law’. This involves transitional legal reforms (e.g. making gender transitioning easier, and legally recognising other gender identities) while also attending to far-reaching structural concerns of poverty, violence, exclusion, and exploitation. Decertification does not rely on these concerns being resolved. However, what decertification means and how it will work will be shaped by the social policy landscape within which its implementation is situated. 

10. In section 9, we set out some possible principles for a law decertifying sex and gender.

Section Nine goes on to state

Legislative principles for the decertification of sex and gender 

The principles that follow provide an example of what decertification could entail if it was introduced as a legal reform. The questions listed identify areas for further discussion and consideration. 

Aims of decertification:

1. To abolish a legal system of certification that treats sex and gender as legally assigned or registered qualities of individuals; 

2. To contribute to the dismantling of hierarchical structures based on gender and sex, that also encode and institutionalise difference; 

3. To support the lives of people whose gender leads them to experience exclusion or other forms of disadvantage; 

4. To contribute to the undoing of social injustices and inequalities more broadly. 

Principles of decertification law: 

1. Legal registration of sex and gender is abolished. 

Sex and gender status would no longer be legally established or assigned (for instance by registering sex on birth certificates). Laws such as the Gender Recognition Act 2004, that provide a mechanism for changing legal sex and gender status, would become redundant. Sex observed at birth could continue to be recorded, in aggregate, for planning and statistical purposes, but would no longer form part of an individual’s legal status. 

Existing legal registration of sex and gender (through birth certification or Gender Recognition Certificate) would no longer carry legal effect. 

2. Introduction of a new ground of gender in equality law. 

Gender remains a legally important term for tackling social subordination, discrimination, violence, and other injustices, including through equality law. Legal use of the concept of gender can also encompass inequalities that relate to forms of embodiment associated with sex. The current grounds of ‘sex’ and ‘gender reassignment’ in the Equality Act 2010 would be merged to form the ground of ‘gender’ as a ‘protected characteristic’ for discrimination, harassment etc. and the public sector equality duty. Recognising gender as a ‘ground’ of inequality and discrimination, i.e. the basis on which inequality and discrimination take place, does not require individuals to be legally assigned to specific gender categories. Employers, service providers, and others also cannot require people to dress or behave differently on grounds of gender. 

Question 

Should gender itself be legally defined? The Equality Act 2010 currently adopts different approaches to different grounds of inequality. In some cases, it works by identifying member classes that together comprise the overall category (e.g. for sex and sexual orientation); in some it describes component elements of the category (e.g. race). As gender’s meaning is in flux, one option is to leave its definition to evolving case-law (as with the ‘protected characteristic’ of religion and philosophical belief). However, since the scope of gender as a legal ground is currently in dispute, an alternative option is to provide a non-exhaustive legislative list of component elements. This could include bodily sex, gender non-conformity, norms and expectations relating to women and men, and social transitioning. 

3. Gender-neutral legal drafting. 

Current ‘gender specific’ terminology includes mother, female, father, male, woman, man, opposite sex, same sex. Building on existing practice, where it is necessary to use pronouns in legislation, gender-neutral pronouns (e.g. they, them, their) should be used except where this leads, or contributes, to structural inequality, other injustices, or to lack of legislative clarity. In contexts where it is legally important to name physical processes with gendered meanings, this can be done without using an explicitly gendered language (e.g. gestational or birth parent rather than mother or woman. This recognises that people other than women also become pregnant). 

4. Legal right to organise gender-specific provision for specific purposes. 

Gender-specific provision, activities, and membership criteria would remain legally valid where this is done to address social subordination, unfairness, violence, or harassment (for instance, women’s domestic violence shelters, women’s sports, community provision for nonbinary and agender young people etc). 

Question 

Should gender-specific provision also be permissible in other circumstances, for instance: • To establish or maintain personal dignity in conditions where mixed gender provision is perceived (by the individual concerned or according to prevailing social norms) as demeaning, embarrassing or uncomfortable (e.g. in certain hospital wards)? • By small, informal organisations or those not in receipt of public or commercial funds, even where these are not intended to address social subordination, unfairness, violence, or harassment (e.g. a men’s tennis club)? 

5. Self-identification. 

Decertification introduces a presumption of self-identification in determining ‘gender’ category membership in line with certain other legal categories, such as sexual orientation and race. 

The term ‘gender-specific’ refers to the use of single or multi-gender categories for provision, activities, and data collection by organisations and individuals, which do not extend to include all gender-based categories. ‘Subordination’ refers to rules, decisions, policies, and practices that sustain, contribute, or lead to socially patterned asymmetries of power in relation to resources, treatment, and regard. 

The term subordination is closely linked to inequality. What it emphasises is the processual character of inequality.  However, it recognises that: a) unlawful discrimination may relate to physical embodiment and/ or others’ perceptions in ways that diverge from self-identification; b) taking up gender-specific opportunities or benefits through affirmative action may also require demonstrated evidence of disadvantage based on gender or a capacity and readiness to represent subordinate and marginalised gender experiences; c) gender-specific provision may draw on alternative or supplementary criteria to self-identification in relation to selecting staff, users, and volunteers (e.g. relevant work experience, suitability). However, evidentiary requirements that undermine a person’s wellbeing and dignity are not acceptable. 

Question 

If alternative or supplementary criteria to self- identification, by an organisation or individual service provider, are legally permitted, should they be subject to regulatory oversight (e.g. by an equality commissioner, specialist tribunal, or court)? 

6. Data collection can continue to use gender- based categories where appropriate, for instance a government survey on homelessness or pay. Questions about gender will normally rely on self-identification. On occasion, other kinds of data may be more useful, e.g. based on service provider or employer perceptions of the gender composition of their users and workplace. Data-gathering in relation to embodied sex should take account of variations in sex development (also referred to as intersex) and the diversity of human bodies. For medical purposes, good practice means asking questions at a higher level of specificity. ‘Are you menstruating?’ rather than: ‘what is your sex?’ since the sex category elicited by this question may not provide useful information on the body that someone has. 

7. Harmonisation. 

Existing laws should be revised to align with the principles for the decertification of sex and gender. Marriage, for instance, should take a single unified form, merging the currently separate legal provisions for ‘same-sex’ and ‘opposite-sex’ marriage. 

8. Levelling up. 

Welfare-related laws that require revision because of decertification should be revised in ways that enhance rather than reduce public provision (e.g. the definition of overcrowding should be extended to two persons of ‘any gender’ who are over ten and share a room, see Housing Act 1985, s. 325). 

9. Recognising plurality. 

Where law uses the terminology of ‘same sex’ and ‘opposite sex’ (e.g. definition of ‘sexual orientation’, Equality Act 2010, s. 12), or assumes that there are two gender statuses, this should be amended to recognise plurality.

Empiricism

'How to Do Empirical Legal Studies without Numbers?' by Sida Liu and Sitao Li in 53 Hong Kong Law Journal comments 

This article addresses this methodological question at a crossroads of empirical legal studies in China. It does not aim to provide a normative defence for the value of qualitative methods. Instead, we demonstrate how a ‘scientific turn’ in the 2010s has made empirical legal research in China almost exclusively about quantitative research and then illustrate how qualitative methods can also benefit from the rise of digital technology. We draw on three recent studies as examples to compare and contrast the methodological challenges and opportunities for doing empirical legal studies without numbers: (1) Ke Li’s book Marriage Unbound as an example of ethnography in combination with archival research; (2) Sitao Li’s article ‘Face-Work in Chinese Routine Criminal Trials’ as an example of trial video observation; and, (3) Di Wang and Sida Liu’s article ‘Performing “Artivism”’ as an example of online ethnography. The discussion shows that, despite the rising popularity of ‘big data’ computational analysis in recent years, quantitative methods are not necessarily more technologically advanced than qualitative ones. Technology-assisted interviews and ethnography can open up many new possibilities in data collection and data analysis, sometimes resulting in more exciting and innovative research. 

Competition

'Meta platforms: How the CJEU leaves competition and data protection authorities with an assignment' by Inge Graef in (2023) Maastricht Journal of European and Comparative Law comments 

Competition authorities can identify a violation of the data protection rules when such a finding is necessary to establish an abuse of dominance under the competition rules. This is the main outcome of the judgment that the Court of Justice of the European Union (CJEU) delivered in Meta Platforms on 4 July 2023. The judgment is the next step in the saga that started with the 2019 competition decision of the Bundeskartellamt (the German Federal Cartel Office) requiring Facebook (now Meta) to refrain from combining user data from different sources beyond its social network. The judgment provides a welcome confirmation that data protection standards can also matter for the interpretation of the competition rules. However, what is more remarkable and less expected is the general framework the CJEU sets out for coordination between competition and data protection authorities building on the duty of sincere cooperation and the clarity with which it evaluates the different legal bases Meta invoked for processing user data. The judgment can become a reference point for assessing the legality of personal data processing by powerful firms, but also leaves competition and data protection authorities with an assignment to explore how to coordinate their work in the future. ... 

In a strongly worded judgment, the Court of Justice of the European Union (CJEU) on 4 July 2023 answered the preliminary questions posed to it by the Higher Regional Court in Düsseldorf about whether the approach of the Bundeskartellamt (the German Federal Cartel Office) in its 2019 competition decision against Facebook (now Meta) was in line with EU law. In that decision, the Bundeskartellamt had held Meta liable under German competition law for imposing unfair terms and conditions as assessed by reference to the standards laid down in the General Data Protection Regulation (GDPR). In its preliminary ruling, the CJEU followed the Opinion of Advocate General Rantos and endorsed the Bundeskartellamt's approach to rely on the data protection rules in the GDPR for establishing an abuse of dominance under competition law. The CJEU also clarified how to interpret the legal bases available under the GDPR for personal data processing activities conducted by a powerful player like Meta and provided a framework to guide competition and data protection authorities in coordinating cases where data protection rules are of relevance for assessing compliance with competition law as well. 

While the strong statements of the CJEU about Meta's personal data processing leave little room for the Higher Regional Court in Düsseldorf to form its own opinion, the CJEU's guidance about how competition authorities should coordinate their investigations with data protection authorities is more open-ended. Even though the CJEU indicates that the Bundeskartellamt in its view met the ‘duty of sincere cooperation’ in the case at hand, future instances may not be as straightforward and require competition and data protection authorities to establish more detailed modes of collaboration based on the pointers provided by the CJEU. This is especially relevant with the entry into force of the Digital Markets Act, which regulates data-related practices that are also covered by the GDPR. 

After giving an overview of the facts and background of the case in Section 2, this case comment summarizes the reasoning of the CJEU in Section 3 and then analyses several aspects of the judgment in Section 4. Section 5 concludes by giving an outlook to the future.

Parking and Drafting

It helps not to be prolix.

In Barrett v City of Cockburn [2023] WASC 384 the Court notes 

 The appellant seeks leave to appeal against judgments of conviction that were entered against him in the Magistrates Court in Fremantle on 24 November 2020 in relation to one charge of parking on a portion of a thoroughfare to which a no parking sign applied, contrary to cl 26(1)(e) of the City of Cockburn Parking & Parking Facilities Law 2007 (Parking Law) (Charge 1), and one charge of driving across a footpath, contrary to s 9.4(b) and s 12.24 of the City of Cockburn (Local Government Act) Local Laws 2000 (Local Laws) (Charge 2). The appellant also seeks leave to appeal against fines of $200 imposed in respect of each of those offences. Finally, the appellant applies for leave to appeal against an order made that he pay the respondent's costs fixed at $8,074. 

An appeal notice was filed in this court on 20 April 2022, almost 16 months after the judgments of conviction were entered and sentence was imposed. Accordingly, the appellant requires an extension of time within which to appeal. 

Although it is very clear that the appellant feels aggrieved, and not only because he was convicted of these offences, it is a great pity that so much time and effort has been expended on a matter that resulted in a total of $400 in fines, particularly when the appellant had the option of dealing with the matter by paying a modified penalty of $100. 

The appeal notice that was filed in this matter, together with enclosures, extends to a total of 30 pages. The purported grounds of appeal are discursive and thus the precise nature of the alleged errors and circumstances that are said to give rise to a conclusion that a miscarriage of justice was occasioned are extremely difficult to identify. It is regrettable that the appellant was also permitted to file and serve a 694-page document, euphemistically referred to as an 'Outline Synopsis' as well as a 60-page 'List of Evidence' after the appeal notice was lodged. These documents have hindered, rather than assisted, the court in resolving this matter. 

In an unsuccessful attempt to bring the appellant's complaints about the primary court's decision into focus, on 6 September 2022 a registrar of this court made orders requiring the appellant to file and serve a 10-page summary of the grounds of appeal, cross-referenced to his submissions. In response the appellant filed a four-page document entitled 'Written Submissions Introduction' and a 10-page document entitled 'Written Submissions', both of which fell well short of achieving their desired purpose. 

In addition to the above documents, and compounding the lack of focus, the appellant filed a 67-page, colour-coded document entitled 'Summary of All Evidence', together with an application in an appeal (and a supporting affidavit) in which the appellant sought an order that he be permitted to tender the material referred to in that document on the appeal. Further, in accordance with additional orders made by the registrar, the appellant filed a 19-page colour-coded document entitled 'Evidence to Written Submission', which appears to constitute an attempt to correlate the evidence referred to in the 'Summary' with the submissions made in the 'Written Submissions', as well as the 694-page synopsis. 

The appeal was listed for directions on 20 February 2023. The purpose of the directions hearing was to see whether the appellant was able to clarify his grounds of appeal, and the contentions made in support of those grounds. It was pointed out to the appellant that because of the sheer volume of material that he sought to rely on there was a risk that the points that he wished to make on the appeal might not be properly identified. 

Following the directions hearing, my associate provided the parties with a document setting out, in summary form, the grounds of appeal that the court understood the appellant wished to argue. The content of that document was based on what had been discussed with the appellant at the directions hearing. After considering the document the appellant then made some amendments, added some comments, returned it to the court and provided a copy to the respondent. The court has proceeded on the basis that this document, a copy of which appears as Annexure 1 to these reasons, sets out all the grounds of appeal that the appellant wishes to rely on (Grounds of Appeal). 

Further orders were also made at the directions hearing on 20 February 2023. Those orders were made because the grounds of appeal complained, in effect, that a miscarriage of justice was occasioned because the appellant was prohibited from being able to use a computer during his trial and because certain documents that were the subject of a summons to produce were not produced at his trial. The orders required the appellant to provide lists of the documents the appellant contended were on his computer to which he was denied access, and the documents he contended had not been provided under summons. The appellant complied with those orders and provided two lists of documents. In keeping with the history of this matter, the documents were provided three weeks after the ordered date, and one of those documents was 26 pages long. 

I have not set out the unfortunate history of this matter simply to criticise an unrepresented appellant. However, and with the benefit of hindsight, the court should not have indulged the appellant. The appellant's approach to this matter has severely hampered my ability to confidently and efficiently identify the key points the appellant wishes to make. ... 

Considering all the matters to which I have referred, and with some misgivings because of the lengthy delay, I have reached the view that it is in the interests of justice to grant the appellant an extension of time within which to appeal against the costs order that was made by the magistrate. Accordingly, I would grant an extension of time within which to appeal against the costs order. I would also grant leave to appeal against that order, allow the appeal against the costs order, set aside the order that the appellant pay the respondent's costs and, adopting a broad-brush approach, order that the appellant pay the respondent's costs in the sum of $6,000.

Defamation

Yesterday's 2nd Reading Speech for the Defamation Amendment Bill 2023 (NSW) seeks to give effect to the Stage 2 reforms program for Australia's uniform defamation laws. 

It states 

 In summary, there are six key reforms: one, a conditional exemption from defamation liability for conduit, caching and storage services, and for search engines in relation to organic search results; two, updates to the mandatory requirements for an offer to make amends for online publications; three, a requirement for courts to consider balancing factors when making preliminary discovery orders against digital intermediaries; four, a new innocent dissemination defence for digital intermediaries, subject to a simple complaints process; five, a specific power for courts to make non-party orders against digital intermediaries to prevent access to defamatory matter online; and six, expanded electronic means by which notices can be served. 

I now turn to the detail of the proposed digital intermediary amendments. Schedule 1 [1] to the bill adds new defined terms to the existing section 4 definitions of the Defamation Act 2005. The new defined terms are fundamental to the operation of the digital intermediary amendments in the bill. I will explain some of the new defined terms later as I outline the substantive amendment provisions to which they relate. However, some of terms apply broadly across the amendment provisions so I will cover those now, starting with the term ''digital intermediary". First, I note that a lot of policy thinking, consultation and refinement sits behind this definition. Digital intermediary, in the context of the publication of digital matter, is defined to mean: … a person, other than an author, originator or poster of the matter, who provides or administers the online service by means of which the matter is published. 

A note makes clear that there may be more than one digital intermediary in relation to the publication of the same digital matter. ''Online service" essentially means any service provided to a person to enable them to use the internet. This includes using the internet to do a range of things, such as sending or receiving content, searching for content, sharing content and interacting with other people. Some examples of an online service are included in a note to the definition to emphasise the range of services covered by the definition. An ''online service" specifically includes a forum created or administered by a person using a facility provided by an internet‑based social media platform that enables users to share content or interact with other users about a topic. 

The term ''digital intermediary" is intended to apply broadly. It was developed to cover the full spectrum of functions considered by the stage two review of the Model Defamation Provisions. It deliberately includes forum administrators. It is also intended to cover any new or emerging functions, given the pace at which technology in this area evolves. Another important aspect of the definition of 'digital intermediary is the exclusions. The definition specifically excludes the author, originator or poster of the matter because they are not intermediaries. "Poster" is defined to mean a person who uses the online service to communicate the matter to one or more other persons. The terms ''author" and ''originator" are not defined. Both terms are used in the existing innocent dissemination offence at section 32. 

For the purpose of the digital intermediary amendments, firstly, ''author" is intended to cover circumstances such as when a person who writes a defamatory statement is not the person who posts it. Secondly, ''originator" is intended to include anyone who plays a role in creating the content. Often they may also be the poster, but in some circumstances they may not—for example, where a person edits and endorses a statement that is drafted and posted by another person. Finally, the term ''digital matter" is defined to mean "matter published in electronic form by means of an online service". This is not intended to affect or limit the meaning of ''matter" in the Act. It is only intended to cover a subset of matter, being digital matter. ... 

I now turn to parts of the bill relating to exemptions from liability for digital intermediaries. Schedule 1 [3] to the bill inserts new division 2A into the Defamation Act. This includes two conditional statutory exemptions from defamation liability that apply to narrow classes of digital intermediaries. In the development of defamation law it has been argued that certain traditional intermediaries, such as telephone lines and postal services, are so passive in the publication process that they are not publishers; indeed, they are mere conduits. The stage two review considered if there are equal passive digital intermediary functions that should have statutory protection from defamation liability for third‑party content. The stage two review concluded that there is a very small group of digital intermediary functions that meet this criteria. 

As a result, schedule 1 [3] establishes a conditional exemption from defamation liability for three specific digital intermediary functions. Firstly, a caching service that stores content temporarily to make onward transmission more efficient will be exempted. For example, this includes files commonly downloaded from a website temporarily and automatically stored to speed up the download time. Secondly, a conduit service whose principal function is to enable users to connect with the internet, send data or receive data will be exempted. This includes internet service providers and email service providers. Thirdly, a storage service whose principal function is to enable users to store content remotely will be exempted. An example is a cloud service provider that enables users to store photos for later retrieval. Proposed new section 10B defines "caching service", "conduit service" and "storage service" and includes examples to illustrate what each definition is intended to cover. 

The policy rationale for this narrow exemption from liability is to recognise the passive role that these digital intermediaries play in the publication process. This does not substantially change the law. These digital intermediaries are generally not the subject of defamation claims and are unlikely to be considered publishers under the Commonwealth test. The intention is to provide clarity and certainty. The exemption would apply irrespective of whether the digital intermediary knew, or ought reasonably to have known, the digital matter was defamatory. Given the breadth of this protection, the exemption only applies very narrowly, and a set of conditions are included to ensure that if an intermediary plays a more active role in a publication—for example, by editing the content—that would make the intermediary ineligible. 

The conditions that apply to the exemption are listed at proposed new section 10C (1) (c). Even where a digital intermediary meets the definition of caching, conduit or storage service, if it played a more active role in relation to the digital matter in question, such as editing or promoting, the exemption would not apply. The stage two review of the Model Defamation Provisions also included careful considerations of the functions performed by search engine providers. Ultimately, it was concluded that a conditional exemption from defamation liability from search engine providers in relation to organic search results is appropriate. 

The policy rationale behind this conclusion is that, firstly, in performing the standard functions, search engine providers have no interest in the content. They simply use an automated process to provide users with access to third‑party content. Secondly, search engine providers are unable to remove content from the internet and can only block access to identified URLs from their search engine. Thirdly, unlike, for example, a social media platform, a search engine provider does not have any relationship with the original author. Fourthly, search engines provide significant public benefit and operate on a massive scale. The exemption for search engine providers applies regardless of whether the search engine provider knew, or ought reasonably to have known, the digital matter was defamatory. 

Given the strength of the protection, the exemption has been designed to apply very narrowly. Firstly, proposed new section 10D (1) provides that the exemption only applies to the publication of digital matter comprised of search results or the publication of digital matter to which the search results provide a hyperlink. ''Search result" is defined in proposed new section 10B. It means a result generated by a search engine that is limited to identifying a webpage on which content is located by reference to one or more of: the title of the webpage, a hyperlink to the webpage, or an extract or an image from the webpage. 

Secondly, proposed new section 10D (1) confines the exemption to publications where the search engine provider's role was limited to providing an automated process for the user to generate the results. An example of a publication that would not be covered by the exemption due to these limitations is an autocomplete suggestion for search terms. Another example is an answer composed by artificial intelligence, such as Bing Chat, in response to a question input by a user. Thirdly, proposed new section 10D (2) provides that sponsored search results are not covered by the exemption. 

The new court power to make orders against non-party digital intermediaries provides for a safeguard where defamatory matter may have been published, even where digital intermediaries qualify for a statutory exemption from liability. The bill will insert new section 39A into the Defamation Act, providing the court with the power to order a digital intermediary that is not a party to proceedings to remove or disable access to defamatory matter online in certain circumstances. The new court power would apply to all digital intermediaries, including those that qualify for the statutory exemptions, meaning that even if a digital intermediary is exempt from liability it will still be possible for orders to be made that the digital intermediary remove access to defamatory material in some circumstances. I will go into further detail about this amendment as I speak about remedies introduced by the bill. 

The bill provides an early determination process for the digital intermediary exemptions. This is at proposed new section 10E, which provides that the judicial officer in defamation proceedings is to determine whether an exemption is established as soon as practicable before the trial starts, unless the judicial officer is satisfied that there are good reasons to postpone the determination to a later state of the proceedings. New section 10E (2) (a) provides a non-exhaustive list of matters that are relevant to this decision. The purpose of the early determination process is to support the policy intent behind the statutory exemptions—namely, to recognise that the role of these digital intermediaries in the publication process is such that they should not be subject to defamation claims. Ideally, the early determination process will mean that time and costs are not expended unnecessarily. The savings and transitional provisions in relation to the statutory exemptions are the same as those for the new innocent dissemination defence. I will briefly outline the intended operation when I speak about the new defence. 

One of the objects of the Defamation Act 2005 is to promote speedy and non-litigious methods of dispute resolution. Part 3 of the Act establishes a procedure to enable parties to settle disputes without the need for expensive litigation by encouraging a publisher to make a reasonable offer to make amends to the aggrieved person. If the aggrieved person does not accept an offer that was reasonable in all the circumstances, the publisher may rely on their offer to make amends as a defence in any subsequent defamation action against them, in accordance with the terms of the Act. 

Section 15 of the Act sets out a number of elements a reasonable offer to make amends must and may include. I will refer to these requirements as the mandatory and discretionary elements of an offer to make amends. The bill includes two proposed amendments to section 15 of the Act. The first is a minor amendment to one of the discretionary elements of an offer to make amends. Section 15 (1A) (b) currently provides that, if the defamatory matter in question has been "published on a website or any other electronically accessible location", an offer to make amends may include "an offer to remove the matter from the website or location". That was added as part of the stage one amendments to accommodate online publications. 

The bill amendments section 15 (1A) (b) to provide that, if the matter is digital matter, an offer to make amends may include an "offer to take access prevention steps in relation to the matter". This amendment broadens the provision by allowing a publisher to offer to remove, block, disable or otherwise prevent access to a matter, and is consistent with wording used in the bill. A more significant amendment is proposed in relation to two of the mandatory elements of an offer to make amends. 

These existing mandatory elements are, firstly, section 15 (1) (d), which provides, relevantly, that an offer to make amends must include "an offer to publish, or join in publishing, a reasonable correction of, or clarification of or additional information about, the matter in question"; and, secondly, section 15 (1) (e), which provides that an offer to make amends must include, "if material containing the matter has been given to someone else by the publisher or with the publisher's knowledge, an offer to take, or join in taking, reasonable steps to tell the other person that the matter is or may be defamatory of the complainant". 

These mandatory elements were not originally designed with digital intermediaries for online publications in mind. They make sense for traditional publications, such hard copy newspapers, that do not remain readily accessible at the click of a button. If the publisher of a newspaper receives a concerns notice about a defamatory statement in a particular edition, they could then offer to publish a correction in a subsequent edition, presumably reaching largely the same audience. However, a digital intermediary may not be able to do these things. For example, a search engine would not be able to publish a reasonable correction for a search result. Also, when defamatory matter is published online, it often stays there. Added to this concern is the ease and speed at which it can be further disseminated to a wider audience. It is understandable then that, for many plaintiffs, their central concern is simply to have the matter removed. 

The bill inserts new section 15 (1B), which updates the operation of these two mandatory elements for digital matter. It provides that, if the matter in question is digital matter, an offer to take access prevention steps may be made instead of or in addition to either or both of the offers mentioned in paragraphs 15 (1) (d) and (e). The purpose of this amendment is to ensure that there is an appropriate avenue for offering to make amends in circumstances where it is not possible or meaningful for online publishers to publish a correction or clarification. It also reflects the kind of remedy that many plaintiffs are seeking in relation to online publications. An important safeguard is that, under the existing section 18, if the defendant seeks to rely on the offer to make amends defence, the court must be satisfied that, in all the circumstances, the offer was reasonable. The offer to make amends changes will apply to offers made after the commencement of the amendments. That is even where the matter is published before the commencement. 

Many originators who post defamatory material online do so using a pseudonym. In order to commence defamation proceedings, the plaintiff must identify and locate the originator. In some recent cases, particularly in the Federal Court, plaintiffs have obtained preliminary discovery orders requiring a digital intermediary to disclose information concerning the originator's identity. Australian courts already can and do consider proportionality, privacy and the risk of abuse of process in exercising the discretion to make preliminary discovery orders. However, there may still be a risk that such orders are abused or have a chilling effect. 

Proposed new section 23A provides that, before making an order for preliminary discovery, the court must take into account the objects of the Act and any privacy, safety or other public interest considerations. This does not provide a new avenue to seek preliminary discovery; it simply applies this requirement over the general rules. While courts already have the discretion to consider these factors, there is value in making consideration of these factors mandatory. This will promote consistency across jurisdictions. It is also in the interests of protecting domestic violence victims and other vulnerable members of society. For example, a person who has published matter online using a pseudonym may fear for their safety. A bad actor could seek a preliminary discovery order on the basis that they want to commence defamation proceedings against the person when the real motive is to find out the person's current location or other contact details. 

Proposed new section 23A would ensure the court takes into account privacy and safety considerations before making a preliminary discovery order requiring a digital intermediary to disclose any identifying information it holds about the person. The savings and transitional provisions for proposed new section 23A are the same as for the power of the court to make non-party orders. I will briefly outline their intended operation when I speak about the power to make non-party orders in a moment.