New corporate players and educational policy: How might the Australian competition and consumer commission help us to understand AI’s associations with educational policy?

Introduction

Any candidate using Facebook can put a campaign message promising one thing in front of one group of voters while simultaneously running an ad with a completely opposite message in front of a different group of voters. The ads themselves are not posted anywhere for the general public to see (this is what’s known as “dark advertising”)…That undermines the very idea of a “marketplace of ideas”, says Ann Ravel,…“The way to have a robust democracy is for people to hear all these ideas and make decisions and discuss,”…“With microtargeting, that is not happening.” (Wong 2018)

Facebook walls and Twitter feeds are sought-after spaces for advertising on a global scale. Political campaigning in the United States 2015 general election saw, on average, 73 million people a month reached with Liberal advertising. That’s approximately 22% of the US population. In comparison, Labour only reached 16 million people in their best month (Ogle, 2017) or 5% of the population. Liberal advertising placed a considerable amount of money into online social media spending and was able to target particular groups of voters with distinct political narratives. A ‘hyperlocal’ messaging was communicated through social media to relate and engage with local issues. This impacted the behaviours of a much-localized population, leading to a shift in political preferences. This global issue is also evident in the United Kingdom. In the Parliamentary Assembly of the Council of Europe, Resolution 2326 (2020), the Parliamentary Assembly considered that ‘data-driven electoral campaigning on social media’, especially ‘dark advertising’ on platforms targeting potential voters, was a growing phenomenon. Suggesting a need for stronger regulations, to safeguard data protection, enable transparency and build public trust, the resolution calls for “a political landscape which is more accountable and less prone to manipulation”. In Australia, Andrejevic et al. (2021) have also discussed this issue, considering consumer education and regulatory options associated with unregulated dark ads on social media. This paper considers the global problem of ‘dark advertising’ in the context of the Australian Educational landscape. It argues that dark advertising targets teachers and is not held to account, leading to new forms of governance in Australian educational systems.

Dark advertising has a growing sociotechnical and economic power that allows market dominance across multiple contexts, including education. The digital age has amplified and exacerbated a host of issues in education. A lack of transparency around educational advertising is one such issue that requires more significant consideration. This paper raises rights-based and legal questions about the appropriate regulatory measures that need to be taken when spending on digital campaigns attempting to persuade teachers to adopt forms of edtech is not reported on, nor is the accumulation of detailed data about them. Previously, advertisers have been exposed as targeting changes in behaviour that influence a country’s voting system (Wong, 2018). It is naive to assume that teacher behaviour is not affected by the business models of social media giants such as Facebook. As such, teachers are working via new educational policy terrains that include potent advertising forms. The role of the teacher is thus considered to be a commercial data point (Arantes, 2020), influenced by the power of advertising money to predict behaviour (Kosinski et al., 2016), which exerts a relatively intangible influence over educational policy (Hogan et al., 2018). This paper is significant as the associations of artificial intelligence (AI) that underpin Dark Advertising with educational policy have primarily been considered in teaching and learning (Ouyang and Jiao, 2021). Teachers’ privacy, agency and autonomy that enables or speaks back to online and targeted advertising, is yet to receive sufficient educational research, and there there are implications associated with policy in terms of consent, privacy, and agency (Rennie et al., 2019; Hakimi et al., 2021; Saunders, 2020).

I consider AI, in the form of Dark Advertising. I propose that Dark Advertising is a new corporate player in educational systems. I focus primarily on K-12 educational systems and the teachers employed to work within them, as commercial apps and platforms in K-12 educational settings across Australia have dramatically grown in recent years. Arantes (2022) states, “Many schools across Australia are now either ‘Google schools’ or ‘Microsoft Schools’ used in and around the classroom, in addition to multitudes of other commercial apps and platforms as digital teaching tools.” As such, K-12 teachers are constantly trialling edtech platforms, leaving cookies and other tracking devices on the teachers’ personal devices, providing data for advertising purposes. A cookie is a text file that collects data from websites such as personal information and de-identified data collected from the teacher (Arantes, 2020). This may include location, device, engagement, and browsing history.

I present a nuanced perspective to the debate about digital teaching tools. The paper does not focus on technology in terms of pedagogy and educational practice. Instead, I consider how advertising for emergent technology may impact teachers’ consumer rights based on reports produced by the Australian Competition and Consumer Commission (ACCC, 2019; ACCC, 2021). I draw on Fourcade and Gordon’s (2020) examination of Statecraft in the Digital Age and various reports from the ACCC, to consider teachers in terms of policy associations with Dark Advertising. By doing so, we can understand tangible associations with policy in the context of Australian education and from an international perspective. I theorize about policy informed by national findings presented by Australia’s Competition and Consumer Commission to highlight and problematize Dark Advertising alongside Australia’s national competition regulator. I do so to reveal the impacts and implications of new forms of governance and their influence on educational policy. As an independent Commonwealth statutory authority, the ACCC’s role is to enforce the Competition and Consumer Act 2010 and associated legislation to, in part, improve consumer welfare and prevent conduct harmful to consumers. By considering the ACCC’s recommendations for consumers, we can consider recommendations for teachers as a consumer of educational technology.

This conceptual paper argues for increased discourse about Dark Advertising as a new corporate player in K-12 educational systems to address consequences associated with insight sales strategies. Consequences considered in this paper focus on associations between ‘Dark Advertising’ the democratic process of education, and teachers’ work, informed consent for teachers’ de-identified data to be used for profit and the potentially invasive data practices associated with privacy. A brief note on the act of Dark Advertising follows, followed by a discussion of the theoretical positioning that underpins the discussion and how it helps understand the changing face of commercial digital data in educational settings.

An insights sales strategy: Dark advertising

Dark Advertising is an ‘Insight’ sales strategy that uses artificially intelligent modelling to predict future events to bolster sales (Yu et al., 2011). It uses predictive insights in search engines, social media, content aggregators, and online advertising to “put a campaign message promising one thing in front of one group of [teachers] while simultaneously running an ad with an opposite message in front of a different group of [teachers] (Wong, 2018). As such, the educational ‘marketplace of ideas’ is modulated and moderated through a commercial lens, diminishing the capacity for democratic knowledge sharing.

Machine interpretable and intangible to the teacher, dark advertising avoids responding to a perceived problem or deficit. Insights drawn from past data predict where problems or deficits may emerge (Michalewicz et al., 2007) to target only those likely to purchase, with a very localized and tailored advertisement. Insight Sales strategies are used by commercial platforms, effectively advancing “a disruptive solution because they target accounts where demand is emerging” (Adamson et al., 2012: 4) rather than where users may perceive demand currently. Not only do insight sales strategies offer a technological solution, but they also appear as though the offer emerged just at the right time. This arguably results in what Williamson (2014) refers to as a mutated form of an educational system governed by a commercially networked form of governance. I consider this paper as a form of theorizing about policy to argue that Dark Advertising is not held to the same scrutiny as those who adhere to Codes of Conduct and Privacy legislation in educational settings.

Insight Sales informed by AI refers to predicting future events to bolster sales, usually based on historical data through artificial intelligence models (Yu et al., 2011). It allows sales forecasting of what, when, and how to best approach a teacher or school. By doing so, it appears to the teacher that the commercial offering that the teacher hadn’t considered, but could be of use, has appeared ‘just in time’. Insight Sales strategies draw on advertising practices that are not personalized and, thus, akin to ‘Dark Web Advertising’ (Gehl, 2021). Here we see users’ anonymized content used so that advertisements appear on general-purpose search engines, but tailored to specific and localized users. Secondly, generating product-level sales forecasts without personalization and presented as AI-informed insights is a well-established practice in multiple contexts. For example, Zunic et al. (2020) describe such insights as a crucial global factor in the retail industry. Zhang et al. (2020) outline how such strategies can optimize the speed and profit of cigarette sales in China. For such insights to occur irrespective of the product, primary data about buyer choices are combined with product production to guide the supply chain via a series of data-driven comparisons.

Insight Sales is a commercial sales strategy that does not respond to a perceived problem or deficit; instead, it predicts where problems or deficits may appear (Adamson et al., 2012), which informs the supply chain. Where ‘prediction and evidence through computation may well be putting faith in a method that, at its heart, a form of speculation’ (Gulson and Webb, 2017: 23), Dark Advertising actualizes such speculation and profits from it. As alluded to in Andrejevic et al. (2021), teachers are working in a space where demand is emerging. Platforms can predict which teachers or schools to target when they are most likely to engage with a yet-to-be-voiced problem so that they can be ‘nudged’ towards purchasing a platform that appears ‘just in time.’ (Yeung, 2016).

Dark Advertising is considered a type of Insight Sales strategy observed as online advertising that is not subject to standard regulatory guidelines (Lindstaedt, 2021). Gehl (2021) suggests that dark advertising is not personalized, as advertisers do not use cookies and other tracking devices. Rather Dark Advertising mitigates these tracking technologies by exploiting the information of those who ‘overshare’. Consider teacher-influencers who develop relationships with audiences (Dousay et al., 2018, March) and produce a great deal of data about themselves (Arantes, 2021).

Dark Advertising can be seen only by those to whom they are delivered, based on predictive insights about when, how, and whom to target for maximum profit (Saunders, 2020). Saunders (2020) suggests that Dark Advertisements are not part of a national conversation, are challenging to track, undermine new ideas in the marketplace, and cannot be fact-checked. While the possibility that such content about education is essentially not available for public inspection, Andrejevic et al. (2021) further highlights that such advertising is also difficult to ‘hold accountable for their content and their pattern of distribution’ (p. 5). The lack of accountability supports the spread of complex forms of discourse that favour commercial platforms. The issue is that insights strategies inform targeting advertising and promulgate dominant discourse captured through engagement metrics (Lilley et al., 2012). Dark Advertising that is distributed to only selected individuals who are likely to engage, by default, makes them unlikely to object. Dark Advertising is considered a new corporate player in educational systems, which results in harmful practices associated with agency, consent, and privacy.

A new positioning for AI-informed associations with policy

In this context, dark advertising provides new digitalized processes of reaching and engaging teachers and intangible forms of education governance that involve multimodal predictions and synchronizing where the perceived need is emerging. Previous research has positioned AI in terms of multimodal predictions, intelligent tutoring systems, and educational data mining (Baker, 2019; Bernacki et al., 2021). AI has also been considered in terms of sociotechnical systems (Williamson, 2017c). For example, digital platforms described as ‘standards-based technical-economic systems’ (Bratton, 2015: 141) are part of a broader mobilization strategy ‘to re-engineer educational institutions in the interest of efficiency and prediction’ (Perrotta, 2020: 53). AI is considered here, in terms of the massive growth of commercially curated educational data and unbalanced governance relationships (Ozga, 2009). AI-informed advertising that targets teachers and educational policy has not received sufficient scrutiny. It is not surprising that scrutiny into the pragmatic implications for teachers’ data, whether at work or working from home, is associated with contemporary forms of sales strategies (Levine, 2013; Williamson, 2017b) is relatively quiet. As Andrejevic et al. (2021) states:

The ability to avoid public scrutiny is, at least for some marketers, an added bonus of targeting – it frees up advertisers to develop and implement strategies that would have been objectionable and caused public and legal backlash where they subjected to shared public scrutiny. (p. 16)

Robust sales strategies without public backlash can drive the uptake of digital tools in the classroom, impacting and modulating the pedagogy and forms of assessment (Sunstein et al., 2017). In this instance, advertising provides signals for the market and shapes social messaging, making accountability harder due to its intangibility and lagging regulatory frameworks (Andrejevic et al., 2021). Consequently, data-driven insights provide commercial platforms with predictive information about teachers, which in turn enables ‘Dark Advertising’ that presents significant educational policy issues.

Further, as many observers have noted, intermediaries reflect a reliance on market mechanisms that redefine the teacher as a consumer (Hogan et al., 2018), in part by including customer-oriented policy documents that seemingly divorce business models from marketplace contexts (Perrotta et al., 2020). Gorur (2017) also notes that many of these intermediaries are hybrids, which Sellar and Gulson (2021) speculate can create new contextually relevant values in education policy. Meaning that academics and private and commercial organizations are merging discourse toward large-scale data-based comparisons to provide forms of governance in education (Perrotta et al., 2020). These insights enable intangible forms of behavioural governance that affect teachers’ consent, agency, and privacy. As such, this new positioning is grounded in an understanding that previous research has demonstrated how heavily educational policy depends on data. Moreover, data-orientated predictive insights are part of the educational landscape. Educational policy has implications such as privacy, consent and agency that we need to reposition.

Teachers consume media at home, at work, and via mobile devices when trialling and using educational apps and platforms; their consumption has enabled contemporary forms of educational governance in need of urgent action. More so, these practices are assumed to be given the green light within policy and educational systems due to the broad and open acceptance of commercialization and commercial digital platforms that use cookies and other tracking devices.

How might the Australian competition and consumer commission help us to understand AI-informed dark advertising’s associations with educational policy?

Reconceptualizing the role of the teacher in a hybrid position of human educator and a consumer data point reveals unbalanced governance relationships (Ozga, 2009) leveraged in Dark Advertising. Drawing on an examination of two reports produced by the Australian Competition and Consumer Commission (ACCC), the lens of Fourcade and Gordon’s (2020) discourse of a dataist state has been used to understand AI-informed Dark Advertising associations with educational policy. The Australian Competition and Consumer Commission (ACCC) is Australia’s national competition regulator and is responsible for enforcing legislation to improve consumer welfare and prevent harmful consumer conduct. As teachers are a hybrid, one part educator and another part consumer, I consider the advice, recommendations, and information from the ACCC (2019, 2021)to be of utmost significance; albeit largely under-considered in educational research. To justify this consideration, I will briefly draw on examples of advice from the ACCC, and then outline the two reports produced by the ACCC.

The ACCC uses the term ‘information asymmetries’ to refer to unbalanced governance relationships, where teachers have less information than the commercial platform. By drawing on the ACCC’s (2019, 2021) findings across Australia, the paper contextualizes their findings in the context of Australian Educational systems to understand associations with educational policy. For example, the ACCC (2019) argues that consumers’ ability to make informed choices is affected by information asymmetries. This could be considered as the teachers being ‘generally not aware of the extent of data that is collected nor how it is collected, used and shared by digital platforms’ (ACCC, 2019: 23) they trial and use in the classroom.

This has been discussed as a form of digital governance informing policy (Williamson, 2018a). An example is where the ACCC (2019: 23) states that the ‘length, complexity and ambiguity of online terms of service and privacy policies’ conceals data practices. This may be because data collection and use information are not available to teachers in educational systems. Alternatively, perhaps, teachers are unaware or unable to understand it due to concealed data practices. For example, concealed data practices “occur when suppliers’ terms provide weak privacy protections for consumers while the extent of those terms, the resultant data practices and the consequences of these data practices, are concealed from consumers” (Kemp, 2019: 11). The ACCC argues such practices are familiar with commercial apps and platforms, stating that the average length of policies for digital platforms contains between 2500 and 4500 words, and takes between 10 and 20 min to read (ACCC, 2019). (Rennie et al., 2019) have considered this process in terms of educational policy. Unreasonable expectations to read and interpret commercial policy are further supported by McDonald and Cranor (2008). They argue that consumers are presented with average 1462 privacy policies per annum, meaning that it would take 76 eight-hour days to read them. Thus teachers are swamped with digital offerings (Rennie et al., 2019) in working conditions where concealed data collection and use, with known information asymmetries, have been normalized.

What follows is a brief outline of the two reports from the ACCC that help us to understand AI-informed Dark Advertising’s associations with Educational Policy. Firstly, the Digital platforms inquiry - final report (ACCC, 2019), and secondly, the Digital advertising services inquiry - interim report (ACCC, 2021).

Why should we consider AI-informed associations with policy in terms of ‘Dark advertising’?

Commercial platforms that are capable of new forms of governance simultaneously facilitate technical elements, commercial profit, and education outcomes. A concrete example of insights sales strategies provided by Andrejevic et al. (2021) demonstrates how advertising content changes through AIs comparison of education level and engagement to nudge behaviour. A ‘nudge’ is a notion borrowed from Political Science. Sunstein et al. (2017: 2) describes nudging as ‘liberty-preserving approaches that steer people in particular directions, but that also allow them to go their way’. Nudging can result in people’s behavioural change (Cheney-Lippold, 2011) in multiple contexts. It is reasonable to assume that K-12 teachers exposed to nudging through their social media are acting in favour of the commercial platform while perceiving that they are acting freely.

Largely under-negotiated by teachers (Hogan et al., 2018), platforms are considered economic-political instruments through which teachers as data points are proprietary (Birch et al., 2020). Nudging and targeted advertising have been the focus of some educational research, seeing targeted advertising as a form of content-linked advertising that relies on proxy indicators (Andrejevic et al., 2021; Wachter, 2019; Saunders, 2020). Insights sales strategies shift away from proxies and links to content. For example, Dark Advertising bypasses proxy measures as it is not content-linked.

Dark Advertising acts by anticipating emergent problems in educational systems through the characterization of teachers’ perceptions of need. The teacher can potentially be a data point to target different consumer groups so that only particular groups see the ads. This enables advertisers the capacity to both target specific teachers as well as experiment on them (Andrejevic et al., 2021). To explain, two teachers can look at the same online advertising after being targeted with an ad. However, with Dark Advertising, each teacher will be exposed to entirely different information on the ad based on their past history.

Variations to the advertising without a human programmer are widely used, thus allowing testing of the ad in real-time (Andrejevic et al., 2021). As the AI ‘learns’ which ad is most likely to be engaged with, it is reasonable that individual teachers are nudged towards engaging with ads that Thaler and Sunstein (2008) argues further consolidate the process of behavioural change. This is seen to be quite different than targeted advertising, as the teacher nor the educational system, had not expressed concern, albeit was targeted with advertising and only the creator of the ad is privy to why the teacher is targeted. In sum, Dark Advertising are not content-linked and afford relatively opaque forces affecting agency (Ouyang and Jiao, 2021) and democratic freedoms (Cheney-Lippold, 2017). By discriminating who sees the ad in the delivery process, Dark Advertising ensures that different information is only visible to its intended target (ACCC, 2019), which also leads to various issues in terms of privacy and consent.

For three reasons, understanding the associations of ‘Dark Advertising’ with policy is considered significant. Firstly, Insight Sales strategies have associations with policy based on the democratic process. Used to leverage the teachers’ data without the teachers’ explicit awareness or understanding (Williamson and Hogan, 2020), Insight Sales strategies don’t nudge teachers in response to a perceived problem; they predict the problem and then nudge teachers towards it. Governed by outputs and outcomes underpinned by policy (Peters, 2009), the educational landscape now includes teachers represented as commercial data points. As a data point, the government targets and monitors them to justify policy (Shore, 2008) and commercial platforms to maximize profit (Cheney-Lippold, 2011). There is a hybridization of action where the teachers’ consumption of educational technology is monitored while trialling apps and delivery, where the human teacher is ‘computationally supervised’ when using apps as part of the pedagogy. This is seen to occur with and through internal and external stakeholders (Buchanan and McPherson, 2019), to position the teacher in a hybrid form, of ‘human educator’ and ‘consumer data point’. This is enabled through the mutation of educational systems towards a decentralized and networked form of governance (Williamson, 2014) where a fleeting, almost ephemeral data-informed advertising economy impacts educational policy (Williamson, 2017a, 2018b; Williamson et al., 2018). Albeit difficult to comprehend, predict, and govern (Andrejevic et al., 2021), Ozga (2009) argues that this governance turn is a shift in strategy that creates new central demands for data. Data that informs insights, enable a relentless comparison of global datasets (Perrotta, 2020), and is a fundamental feature of educational governance (Ozga, 2009), as emulated via Insight Sales strategies.

Secondly, the lack of informed consent raises various issues beyond perpetuating dominant discourse, which Thompson and Cook (2013) note is fundamentally driven by deficit rhetoric. Programmed into policy through de-identified data (Williamson, 2014, 2015), the data-informed policy is based on commercial platforms governing ‘what’ the problem is, where teachers’ profitable deficits are emerging, and informing policy with seemingly objective datasets (Lingard, 2018). Teachers’ negotiation of commercial policy is complex due to asymmetries in information about consent and how de-identified data is used. No teacher would willingly consent to a commercial platform to intangibly profit from an erroneous discourse of their perceived deficits. The ACCC (2019: 23) reports on what is referred to as a ‘considerable imbalance in bargaining power between digital platforms and consumers’. Such imbalances limit the ability of teachers to have the well-informed choice, deemed essential to freely give consent in terms of the collection, use, and disclosure of their data. Informed consent is seen to require the teacher to consent for their personal data to be collected by cookies and used by a commercial platform for profit. However, to give consent for their de-identified data is not part of this current definition (Mai, 2019). Further, to give consent for de-identified data to be collected and used, the teacher must have explained to them what data is collected, what it is being used for, how it will be stored, who has access to it, and when it will be deleted (Royakkers et al., 2018). This is often provided by only a ‘tick box’ opt-in (Rennie et al., 2019), which the ACCC (2019) argues does not give meaningful consent. This argument is supported by Obar and Oeldorf-Hirsch (2016), who suggests that there are deep flaws in assuming people can interpret legal policy associated with ‘opting in’ via a tick box.

Thirdly, some Insight Sales strategies rely on potentially invasive data practices (ACCC, 2021) that may contain sensitive personal data which can re-identify a teacher (Culnane and Leins, 2020). As such, there may be concerns associated with privacy. Rennie et al. (2019), in an exploration of privacy and app use in Australian primary schools, found that 60% of teachers did not explicitly consider privacy. Further, privacy was not considered in terms of local policy and legislation. Rennie et al. (2019) state ‘The apps that we found schools to be using do not necessarily reference Australian law in their privacy statements, instead referring to the laws and frameworks of the countries where they were created (if at all)’ (p 8). Acknowledging that in Australia, apps used by teachers need to comply with state or territory privacy laws, the Privacy Act and the Australian Privacy Principles (Rennie et al., 2019), multiple authors argue that there are numerous privacy issues to be considered (Rennie et al., 2019; ACCC, 2019; Obar and OeldorfHirsch, 2016).

Further research: codes of conduct and legislation

The Australian Competition and Consumer Commission helps us to understand Insights Sales in the context of educational policy and new forms of governance. I will finish by considering to what extent Insights Sales strategies may circumnavigate established Codes of Conducts and Privacy legislation.

Maxwell (2017) discusses codes of Professional Conduct for future teachers in Australia, England, the Netherlands, and the USA. Discussing how such codes are ‘consciousness raising about how codes of ethics are used to judge allegations of teacher misconduct’ (324), Maxwell (2017) stresses that ethical conduct in teaching and learning is paramount. Forster (2012) in a comparative review of Australian state and territory codes of ethics and conduct, argues that teachers are expected to share values and promote moral capacities and sensitivity.

Drawing on the ACCC final report (2019), there is a recommendation that commercial platforms should implement their own code of conduct. This is arguably supported by Campbell (2006), who notes that teachers must act as moral agents and values educators. For example, Google has a Code of Conduct that is ‘measured against the highest possible standards of ethical business conduct’ (Google, n.d.) and Ambassadors for social media edtech tool Edmodo (Edmodo.com) have a Code of Conduct that asks them to ‘Advocate for Edmodo’ Edmodo (May 21, 2021). The Edmodo Code of Conduct states that ambassadors must ‘Be positive and optimistic when talking about Edmodo, whether you are at a conference, on social media, or on the platform. Teachers learn best from other teachers, and your enthusiasm and support make a huge impact on Edmodo’s future’ (Edmodo, May 21, 2021). Thus the codes of conduct are correlated to business drivers rather than shared values, moral capacities, and educational misconduct. The ACCC (2019) notes that past experience, where commercial platforms have been expected to develop, implement and adhere to their own code of conduct, has not worked. What can we do as educational researchers?

The ACCC’s (2019) report proposes amendments to Australian privacy law to protect the long-term interest of consumers whilst maintaining the data-driven markets that underpin the digital economy. As a result, scrutiny of Dark Advertising (as discussed in a dataist state in education) alongside the ACCC’s reports (2019, 2021), leans more towards issues of legislation than codes of conduct. Among other factors, as part of the recommendations provided by the ACCC, there is a call to update the definition of personal information. The call seeks to change the definition of personal information in the Privacy Act ‘to clarify that it captures technical data such as IP addresses, device identifiers, location data, and any other online identifiers that may be used to identify an individual’ (ACCC, 2019: 456). This call is supported by the Australian Government (Frydenberg et al., 2019), and justification for such a change is increasingly receiving research (Culnane et al., 2017; Culnane and Leins, 2020). By drawing on the ACCC’s reports (2019, 2021), we can see that scrutiny of teachers as a data point requires educational policymakers to consider whether the collection and use of teachers’ personal information, whether de-identified or not, aligns with suggested changes to the Australian Privacy Act (Australian Government, 1988).

Scrutinizing the data collected about teachers is paramount, particularly if educational systems encourage teachers to use commercial apps and platforms that collect and use teacher data. Such scrutiny requires urgent considerations, as these ‘just in time’ offerings, based on historical data, remain intangible to the teachers and schools impacted by them. In sum, and according to the findings of the ACCC, Dark Advertising, as a new corporate player in education must be more widely examined. I call for further research in this space and strategies to address such considerations.

Conclusion

Dark Advertising in education is a new corporate player with associations to educational policy. Only one of many processes, under the umbrella of insights sales, it is largely machine-interpretable and intangible to the teacher and school, Darks ads have been shown to demand greater scrutiny. Using the teacher as a data point and with cookies on teachers’ personal devices accepted mainly as ‘part of the workplace’, insight sale strategies’ power in education is hugely under-scrutinized.

The reconceptualization of the role of the teacher as a data point enabled the intangibility of Insight Sales, Dark Advertising, and the commercial reconstruction of educational governance to be aligned to the ACCC’s findings. As Insight Sales predict a need or want for technical elements based on the historical demands of an educational context, it is by making the ‘fluid legible’ (Fourcade and Gordon, 2020: 86), that this paper adds to the body of knowledge in this space. This paper has demonstrated why consideration of Dark Advertising is essential. Without greater scrutiny, the role of the teacher as a data will continue to be leveraged for Dark Advertising, not addressing its associations with educational policy.