Translating Principles into Practices of Digital Ethics: Five Risks of Being Unethical

“In environmental ethics, greenwashing (Delmas and Burbano 2011) is the malpractice of a private or public actor seeking to appear greener, more sustainable, or ecologically friendlier than it actually is.”

“In the ethics of AI, see for example (Wiener1960; Samuel 1960).”

“The ethics of artificial intelligence in general has received a great deal of attention, by researchers (e.g., Boddington 2017; Floridi 2019; Jobin et al. 2019; Whittaker et al. 2018; Winfield and Jirotka 2018) All of these efforts principally focus on data (involving issues such as informed…”

“The counterpoint to ethics dumping is ethics shirking : what is not done to protect people when harms emerge beyond what is demanded from legal or regulatory frameworks (Floridi, 2019).”

“It is, therefore, a necessary building block in the creation of an environment that fosters ethical, responsible, and beneficial ML, especially as it also indicates the possibility of a time when the distractive risk of ethics shopping 13 (Floridi 2019c) will be lessened.”

“AI companies try to convince people that their behaviors, products, and services are ethical by “Ethics Bluewashing”, but it is not the case [15].”

“…al., 2016; Bro˙zek et al., 2024), legal and regulatory frameworks (Selbst and Powles, 2017; Kaminski, 2019), ethical principles (IEEE Global Initiative on Ethics of Autonomous Intelligent Systems, 2019; Floridi, 2021), and multi-stakeholder collaboration (Lehr and Ohm, 2017; Young et al., 2019b).”

“(Floridi 2019b, 187).”

“Floridi [20] has suggested that developers of AI may conduct ethics shopping, borrowing liberally from different frameworks to arrive at a set of easy-to-implement norms.”

“Floridi [120] asserted that the plethora of different AI frameworks that have been proposed over the years have created confusion and inconsistency among the AI community in terms of the complexity and intricacy of adhering to specific AI ethical situations and contexts.”

“…systems: bias amplification, data privacy, lack of transparency, human oversight, accountability, etc. (Royal Society (Great Britain), 2017; Fjeld et al., 2020; Campolo et al., 2017; Floridi et al., 2018; Floridi, 2019, 2021; High-level expert group on artificial intelligence (HLEG, AI), 2019).”

“We are up against a system that has veritably mastered ethics shopping , ethics bluewashing , ethics lobbying , ethics dumping , and ethics shirking [33].”

“When something (ethical tools, methods or guidelines) is too flexible it does little to protect against the risks of ethics shopping and ethics washing (Floridi, 2019b).”

“The principles may also be misused in the prosecution of personal or corporate interests (Floridi, 2019).”

“This has created a scenario in which those producing, purchasing, or using AI systems can ‘ethics shop,’ ‘ethics wash,’ and ‘ethics dump’ without fear of facing any real consequences (Floridi 2019).”

“This then leads to ethics washing and its cousins: ethics shopping, ethics shirking, etc. (Floridi, 2019; Greene et al., 2019; Wagner, 2018).”

“…ethical principles from different sources to justify some pre-existing behaviour; ‘ethics bluewashing’, i.e. making unsubstantiated claims about ADMS to appear more ethical than one is; and ‘ethics lobbying’, i.e., exploiting ethics to delay or avoid good and necessary legislation (Floridi 2019).”

“At the time of writing, nothing is compulsory in terms of regulation, AI governance, ethics, and legal compliance trail the rate of AI innovation and implementation in digital society (Floridi 2019; Roitblat 2020).”

“Machinewashing can represent a profound nonmarket strategy involving the full-level lobbying spectrum directed at the government, civil society, and academia (Bietti, 2020; Floridi, 2019; Ochigame, 2019; Rességuier & Rodrigues, 2020).”

“No doubt, concerns for ‘ethics washing’ [50, 51] may arise when these two features are pursued in tandem, as we discuss further below.”

“It is what I have called the Green and the Blue human project (Floridi, 2019b; Floridi & Nobre, 2020), which the EU can and should promote in the world.”

“This is akin to what Floridi [71] has labelled ‘ethics dumping’, i.”

“But if the top-down approach to algorithm ethics is too vague, a pragmatic one might be too flexible (Floridi, 2019).”

“Unless monitored carefully, this patchwork of protections potentially enables companies to ethics shop or ethics dump, that is, to choose the most convenient and least restrictive location to conduct a specific task in the algorithm development pipeline [19].”

“In addition to that, ethics can be used for marketing purposes (Floridi, 2019; Wagner, 2018).”

“Floridi (2021) emphasises the importance of developing such ethical frameworks and converting them into practical tools, while others question the difficulty of applying them (Hagendorff, 2020), the lack of empirical evidence about their influence (McNamara et al., 2018), the lack of consideration…”

“The word ‘principles’ is used in the title of most of these document (e.g. Google, 5 Microsoft, 6 Telefonica 7 ), and in the literature (e.g., Fjeld et al., 2020; Floridi, 2019).”

“Consider, for instance, the recent criticism that the AI industry utilizes soft governance merely for purposes of “ ethics washing ” and for delaying regulation (Butcher & Beridze, 2019; Floridi, 2019b).”

“5 See also, Floridi (2019).”

“…on ethical principles as non-binding guidelines (i.e., soft law) and self-regulatory commitments may lead to ethics shopping (Floridi, 2019a, b), that is, the malpractice of selecting ethical principles that align with existing business practices, and justifying them a posteriori (Floridi, 2019b).”

“Over 70 recommendations (Floridi 2019) were identified based on a combination of sources from academia, industry, and government (Morley et al. 2020).”

“…2020; Estrada 2020; Mittelstadt 2019b; Hagerty and Rubinov 2019; LaGrandeur 2020; Lauer 2020; Sloane 2018), as well as the contextual applicabil‑ ity in some areas of these ethical guidelines (Ryan and Stahl 2020; Morley et al. 2020; Hickok 2020; Floridi 2019; ÓhÉigeartaigh et al. 2020).”

“…including the challenges of bias in datasets, the protection of data privacy, the imperative for unbiased assumptions and decision-making, and the pursuit of transparency throughout the processes of data collection, analytical processing, model development, and interpretive practices [63, 93, 162].”

“…specific regulatory measures and oversight can lead to practices such as ethics washing (introducing superficial measures), ethics shopping (choosing ethical frameworks that justify actions a posteriori) and ethics lobbying (exploiting digital ethics to delay regulatory measures) (Floridi, 2019).”

“practices or laws present in a single region, being endorsed not by internationally recognized standardization bodies but by other stakeholders with more "appealing" code of ethics and standards [55, 142].”

“, & Floridi, L. (2019). The debate on the ethics of AI in health care: A reconstruction and critical review.”

“However, implementing this emerging technology in this fields brings social, economic, and primarily ethical implications (Benefo et al. 2022; Floridi 2019; Jia and Zhang 2021; Martin 2019).”

“This has resulted in significant inconsistencies in the ethical frameworks governing the acceptable use of these digital data (Floridi, 2019).”

“How to apply ethics to the development of ML is an open question that can be solved in a multitude of different ways at different scales and in different contexts (Floridi, 2019a).”

“In the latter case, the proliferation of diverse sets of principles can create a real “market for principles” [73], where organizations and stakeholders are encouraged to choose for those principles that most justify and do not hinder their current behaviours.”

“Although much research has focussed on ethical issues related to the outcomes of AI systems (Floridi 2019; Hagendorff 2020; Jobin et al. 2019), another key consideration is the ethics of how AI systems are produced.”

“Guidelines for Responsible AI do have an important role as soft regulation on this path, but they are often either too rigid or too vague and run a considerable risk of ethics shopping and bluewashing (Floridi, 2019).”

“Explainable systems could be implemented as a superficial ethical measure to avoid necessary regulation [38].”

“It also has a clear impact on the historically discriminated groups (Alessina and Glaeser 2013) and it is closely related to gender discrimination in capitalist development (Folbre 2021).”

“Moreover, especially for companies, there could be reasons leading to ignoring part of the principles: among these, the Ethics Bluewashing [9] (the unethical behavior of making unsupported or false claims about, or taking insufficient action to support, the ethical ideals and advantages of digital…”

“Bioethics is the applied ethics domain that most closely parallels digital ethics and the five bioethical principles in dealing with new types of specialists, patients, and conditional axioms are not exhaustive and are inextricably linked to digital morals (Floridi 2013; Floridi et al. 2019).”

“Although the ethics guidelines published by AI HLEG are an example of soft governance, they presuppose and are aligned with the EU legislation (Floridi 2019a).”

“A great range of ethical principles, codes, guidelines, or frameworks have been introduced very recently [8].”

“But such abstract value-definitions risk becoming the target of a common criticism of AI ethics principles for being superficial “ethics bluewashing” (Floridi 2019, 187), and they are to a large extent not precise enough to be accounted for in concrete algorithmic design (Stray 2023).”

“However, considerable doubts remain that corporate communication about ethical AI does not match daily business conduct [6, 30, 93], and, similarly to the long-established greenwashing concept, corporate actions are seen as a mere means to attain legitimacy [97, 100, 120].”

“This practice is called “ethics washing” [25, 151, 171] or “bluewashing” [56] analog to “greenwashing” [50].”

“This moral distance problem could also be related to two of the unethical risks of the translations of AI principles into practices, identified by Floridi (2019).”

“Explainable Artificial Intelligence (XAI) is formed by a set of techniques that allow a wide range of stakeholders to understand why or how a decision was reached by an algorithmic system [61, 164].”

“This possibility would substantiate the fears Floridi (2019) has voiced about phenomena like ‘ethics shopping’ and ‘ethics blue washing’.”

“Floridi [20] identifies other tactics, such as imposing self-regulation while lobbying against new legal norms (“ethics-lobbying”),engaging only with particular ethical principles that retrospectively justify a company’s behavior (“ethics-shopping”); employing PR activity to address an ethical…”

“…policy documents on AI is that the main focus has so far been on the formulation of principles (Mittelstadt, 2019), and there has been less focus on how users and designers of AI can apply these principles to shape the use and development of AI (for an exception, see for instance Floridi, 2019).”

“Aside from concerns about the enforceability of ethical design, scholars point to ethics washing, or the signalling of ethical concern without corresponding action (Bietti, 2020; Floridi, 2019; Wagner, 2018).”

“This transparency asymmetry suggests a weaker form of “safety-washing”, where safety and ethics frameworks remain high-level and the empirical evidence required to rigorously assess risk is selectively disclosed [18, 43, 104].”

“18 Though they may follow organizational ‘best practices’ by establishing ethics boards and collecting feedback from a wide range of stakeholders, an actual commitment to moral principles may mean doing more than what philosophers are allowed to do within corporate settings.”

“All told, the ethics of AI has garnered a great deal of attention from scholars around the world (Floridi 2019; Jobin, Ienca, and Vayena 2019) as well as leading initiatives emerging in predominantly Western contexts.”

“Similarly, such an approach may be used for the regulation of practices of ethics dumping, washing, shopping, lobbying and shirking (Floridi 2019).”

“The use of certification standards as accountability mechanisms in AI has been criticized for potentially promoting an instrumen-talization of ethical values without necessarily eliminating irre-sponsible and unethical behavior [38, 99].”

“Building legitimacy and dependability in aI systems is facilitated by prioritising trust in the aI design (floridi, 2019).”

“Some firms may publish acceptable use policies as a type of public relations statement to demonstrate they are responsible organizations, as firms incur no costs for doing so if they do not invest in enforcement [52].”

“This is an issue that is also noted by Floridi, who refers to this as ethics lobbying [69], as well as by Ress é guier and Rodrigues [70] and Benkler [71], the latter stressing the danger of letting the industry alone to shape the future of AI by writing the rules.”

“Policymakers will also have to recognise the limitations of industry assurance mechanisms, and where market forces could create conflicts of interest leading to unethical practices (Floridi 2019).”

“Floridi (2019) identifies risks in the transition from what to how that include: digital ethics shopping, “bluewashing” (i.e., digital greenwashing), lobbying, ethics dumping (outsourcing to other actors), and shirking.”