The original text of this speech was delivered by Andrew Low, Partner, at the Competition Law Conference in May 2025 at the Sheraton Grand Sydney Hyde Park as a comment to the Chair of the ACCC’s opening remarks.
Thank you to the Chair, Gina Cass-Gottlieb, for sharing the regulator’s perspective on the Digital Platform Services Inquiry (DPSI) and thank you to Chris and Christine for once again organising this conference. Likewise, I would like to acknowledge the Traditional Custodians of the land we are meeting on, the Gadigal people of the Eora nation. I pay my respects to their Elders past and present.
It’s a privilege to share my commentary on the ACCC’s DPSI, marking the end of a substantial five-year endeavour to inquire into and analyse digital platform markets in Australia. Before I start, I wanted to congratulate Kate Reader, Morag Bond and the ACCC team for the significant work and commitment to the DPSI, and I also wanted to acknowledge the significant investment of time and money by the business community and advisers (many of whom are in this room) throughout the DPSI.
To appreciate the end, it is often useful to reflect briefly on the beginnings
We are awaiting the publication by Treasury of the ACCC’s final DPSI report. To appreciate the end of the DPSI, it is useful to reflect specifically on its genesis and development over the five years.
The DPSI commenced in 2020 in the wake of the ACCC’s first 18-month Digital Platforms Inquiry (DPI) in 2018-2019. As many may recall, the DPI and the introduction of the News Media Bargaining Code, were the quid pro quo for then Senator Xenophon’s support for the introduction of new media laws. It had a very specific intent – it was primarily directed to understanding the impact of large digital platforms on news media and the quality of journalism in Australia.
During the course of the DPI, in March 2018, the Cambridge Analytica scandal came to public attention and led to a heightened sense globally that regulators did not have a handle on the way digital disruption had reshaped various market structures for better or worse. This led to the broader five-year inquiry of the DPSI.
After the DPSI kicked off, Australians were launched directly into COVID-19 with the first national lockdown on 29 March 2020. This reshaped the digital economy in many ways, accelerating the world’s transition from offline to online:
Virtual meeting companies such as Zoom saw their revenue skyrocket. In the 12 months ending 31 January 2021, Zoom’s revenue increased more than 300% to USD$2.65 billion.
Stores went cashless, accelerating the decline of cash and accelerating the adoption of digital contactless payments such as ‘tap and pay’ mobile wallets.
There was a boom in retail platforms, as small businesses increasingly relied on these platforms to service customers remotely.
In March 2023, mid-way through the DPSI, GPT-4 was made publicly available. It was subsequently replaced in 2025 via OpenAI’s API. The disruptive and reshaping effects of generative AI were significant for digital platforms. AI agents are now integrated on our phones, applications and search engines – and can generate synthetic data to train AI models, breaking the need for large databanks.
Reflections on the digital economy as a whole
One way to reflect on these transformations over the past five years is that the regulator’s goal is to contribute to a regulatory environment that is coherent, fit-for-purpose and responsive to emerging challenges, as it has done since 1974. Another reflection is that any new regulatory regime needs to respond to a fast-moving, opaque and complex digital platforms market – and that enforcement through litigation, which can be lengthy and often provides only limited remedies, may not be adequate. Both are fair reflections.
What’s particularly interesting to me about these developments is that, while the DPSI is thorough and illuminating, we should not assume that all the homework on understanding digital markets is ‘done and dusted’. The issues of tomorrow are not necessarily the issues addressed by the DPSI. For example, the DPSI Ministerial Direction did not specifically consider some critical and emerging issues that occurred during the five years, such as digital payments. Further, the DPSI factual findings itself showed that the future direction of digital markets cannot be predicted with a high degree of certainty. In the DPSI Ninth Interim Report, the views about search dominance became increasingly tentative in the wake of generative AI.
I’m not saying we necessarily need another inquiry, but policymakers need a better way to use the DPSI’s factual findings and recommendations to inform a process for regular and timely information and feedback loops into any digital reform. As the Furman Report recognised, “It is vital that some flexibility to agree revisions is retained in such fast-moving markets where innovation is so central to companies’ business models”.
What you (don’t know) you don’t know can kill you
One of the prevailing challenges for digital platform markets is that there may be a lack of transparency in the operations and terms between business-to-business and business-to-consumer relationships, posing difficulties for traditional competition and consumer law enforcement.
In 2003, Dr Laura Nathanson wrote a book ‘What You Don’t Know Can Kill You: A Physician’s Radical Guide to Conquering the Obstacles to Excellent Medical Care’, a title playing on the old proverb ‘what you don’t know can’t hurt you’. It speaks to navigating the health care system by empowering patients to identify signs of misdiagnosis, misleading analysis, miscommunication and unknown unknowns.
I think this is an interesting theme for commenting on the regulatory challenge and response. After five-years of the inquiry, what do we think we don’t know, we don’t know?
A couple of observations.
First, under the reforms, when it comes to case selection for services to be ex ante regulated, how effective is the system for identifying services that may require intervention?
The Treasury’s proposal paper (December 2024) envisages a minister or ACCC initiation for a designation investigation. An important element to discovering services that may need intervention is the mechanisms by which regulators receive complaints. Have we given sufficient attention to practices that could distort this?
On 30 January 2022, in a letter to the U.S. Securities and Exchange Commission, a complaint was made that NDAs were being used by Apple against their employees to restrain the disclosure of unlawful acts. It was reported in 2021 that Apple includes gag order clauses to prevent disclosures around digital payments. ‘Gag’ practices in B2B relationships prohibiting regulator disclosures can be more insidious as they are inherently opaque. They’re not observable by the consumer. Gag practices also make it very difficult for regulators to receive complaints from corporate victims. Do we need to think about how to help businesses stand up for themselves or better inform regulators?
Second, what are the unintended consequences for the proposed reforms or codes and how do we study and verify them? Let’s take a simple example of ‘choice screens’. Choice screens were proposed as a possible solution to address exclusive pre-installation arrangements and harms for dominant providers. But experience shows consumer ‘choice screens’, once mandated, tend to inhibit competition by entrenching the prevailing dominant provider (that is, entrenching the existing consumer bias). Thinking broader – do we understand the full impacts of intervention, both in terms of competition, but also consumer and commercial investment? What if we ‘get things wrong’ – is there a sufficient off-ramp and a willingness to use it?
Regulatory challenges – is ex ante regulation the panacea we are hoping it would be for outcomes in digital markets?
The ACCC and Chair have consistently referred to the complexities of grappling with competition and consumer issues in the digital space and the insufficiency of existing investigation and enforcement methods to address the harms identified in the DPSI. There are merits to these concerns. I would make two observations.
First, what’s interesting to me is the importance of complementary enforcement action with ex ante regulation.
Ex ante and sector specific codes could suffer the same issues of delay, at the end of the day, while clearer and more targeted, also slow in design and slow in enforcement. For example, in Europe, the process of declaration of gatekeepers under the Digital Markets Act (DMA) has taken significant time to deliver outcomes for consumers. A number of companies have appealed the designation of gatekeepers or services in some cases taking more than a year and a half since the initial designation (related in part to issues associated with the regulatory design). While fines were issued on some platforms last month for alleged breaches of the DMA, at the time of this address some platforms have reportedly indicated that they will appeal the fines. The ACCC Chair has also previously acknowledged that ‘legal action by competition authorities around the globe has not led to a sustained change in the conduct of the largest digital platforms or their core business models in digital markets.’
There is evidence that enforcement actions can drive outcomes sooner and in a more targeted way, than ex ante regulation. For example, the undertakings provided by Telstra, Optus and TPG prevent them from renewing or entering into exclusive agreements with Google for the pre-installation of its search services as the default on their devices. There are other international examples. In Germany, Lex Apple Pay (ex ante regulation for the Apple Pay NFC chip) passed in 2021, but did not deliver outcomes as there was contest on what constituted a reasonable price associated with access. The DMA clearly intended to cover access to the NFC to enable mobile wallet payments, but has not delivered any real outcomes for such access to date. However, the enforceable undertakings accepted by the EU on 11 July 2024 opened access to the NFC in the context of resolving an enforcement investigation by the European Commission into Apple Pay – before the ex ante rules were passed. That’s not to say that the threat of regulation did not play a role in the resolution of the enforcement action – but they are complementary.
In Japan, while the Mobile Software Competition Act is not in effect and Apple and Google have been designated, the Japan Fair Trade Commission at the same time has issued cease and desist orders to Google for violations of the Antimonopoly Act. The complementary approach appears to have merit.
There may be a tendency overseas for ex ante regulations to detract from traditional enforcement that allows for targeted and specific focus on ‘bad conduct’ – that may in instances deliver outcomes sooner than waiting for and then relying on, potentially lengthy ex ante regulation.
There may be matters in digital markets that are not a regulatory grey zone. Pre-occupation of resources and time on ex ante regulations should not crowd-out much needed enforcement focus and attention on specific cases of harm using the available tools we have, such as the (new) s 46 provisions refashioned in 2017.
Second, should ex ante regulation go hand in hand with enabling regulation that allows or facilitates market actors to innovate themselves out of harm?
Historically, markets have been the best solution for solving digital issues. The Kazaa music wars were ultimately put to bed by low-cost streaming platforms like Spotify and Apple Music. Microsoft’s Internet Explorer, the focus of the US v Microsoft competition litigation, was not resolved by litigation – but rather the rise of newer and faster browsers such as Firefox and Chrome.
This approach was alluded to by a number of submissions to the DPSI and the DPSI interim reports itself. For example, the DPSI Ninth Interim Report and various submissions pointed to the development of generative AI as the leading potential to disrupt general search markets and the perceived entrenched dominance of Google. To address harms associated with search, is there a case for ensuring our regulatory settings are open to and not impeding the development and diffusion of generative AI solutions, including by having settings that facilitate investment and adoption of AI? Further, in the DPSI Eighth Interim Report, the issue of extensive third-party data collection leading to rising fraud and identity theft online was identified and it was recognised that the Australian Government Digital ID system could reduce the risk of raw personal data being transferred through government online identity verification.
It's all about the facts
The DPSI reports to date are an extensive expression of the facts. These facts tell a story and there are two I find particularly interesting:
First, regulatory solutions should be discerning on a case-by-case basis. For example, the facts uncovered by the DPSI include that no two ecosystems are the same.
Second, many business models and practices are not inherently good or bad – they have elements of both. For example, data access and broking are beneficial to the economy, cost-efficient functioning of business and intelligent innovation – but at the same time raise consumer vulnerability and competition concerns.
The Chair has also acknowledged the significant benefits delivered by digital innovation. We should not underestimate the promise for digital innovation. In a book titled ‘Technology and the Rise of Great Powers’ by Jeff Ding of George Washington University, Ding describes how general-purpose technologies (GPTs) such as AI and their diffusion are closely tied with the economic power of a nation – and that the failure to promote diffusion of new GPTs such as AI could lead to an overall decline in economies.
A challenge for any digital ex ante regime is to walk the line between preserving the benefits, while mitigating the harms – or ‘separating the sheep from the goats’ – and being transparent around the framework of analysis being applied to assess harms and benefits (on an economy-wide basis) and to be fair, impartial and open to all facts applied in that framework.
One of the main criticisms of the EU’s DMA is that the obligations require actions for things that businesses are not involved in.
Concluding remarks
In conclusion, is the DPSI the stepping stone directly to designing reform? Or does the DPSI highlight the importance of care and clear assessment frameworks, when assessing when intervention is warranted? Perhaps these two paths are not mutually exclusive.
The issues are complex. The DPSI identifies the nuances between developments that are positive, yet contain negative elements. The DPSI required the ACCC to undertake a broad inquiry but at a specific point in time – a period during which many things have since changed.