Ethics Principles, Challenges, Cases, and Strategies

permalink: "justice"
- - Specific ethical concepts arising in the context of technology
  - And for education

to add here

tbc

Justice involves concern that the benefits and burdens of research must be fairly distributed; those who bear the greatest burden should benefit from the outcomes of research, and established inequities should be considered in making judgements regarding research foci and participation.

Research that is just involves:

"taking into account the scope and objectives of the proposed research, the selection, exclusion and inclusion of categories of research participants is fair, and is accurately described in the results of the research
the process of recruiting participants is fair
there is no unfair burden of participation in research on particular groups
there is fair distribution of the benefits of participation in research
there is no exploitation of participants in the conduct of research
there is fair access to the benefits of research.
Research outcomes should be made accessible to research participants in a way that is timely and clear." (NS 1.4-1.5).

“ Treat all individuals equally and protect social equity Use digital technologies as an essential support for the protection of fair and equal treatment under the law Prioritise social welfare, public interest, and the consideration of the social and ethical impacts of innovation in determining the legitimacy and desirability of AI technologies Use AI to empower and to advance the interests and well-being of as many individuals as possible Think big-picture about the wider impacts of the AI technologies you are conceiving and developing. Think about the ramifications of their effects and externalities for others around the globe, for future generations, and for the biosphere as a whole” (Leslie, 2019, p. 11)

recztovrpeflfq0aw

“The development, deployment and use of AI systems must be fair. While we acknowledge that there are many different interpretations of fairness, we believe that fairness has both a substantive and a procedural dimension. The substantive dimension implies a commitment to: ensuring equal and just distribution of both benefits and costs, and ensuring that individuals and groups are free from unfair bias, discrimination and stigmatisation. If unfair biases can be avoided, AI systems could even increase societal fairness. Equal opportunity in terms of access to education, goods, services and technology should also be fostered. Moreover, the use of AI systems should never lead to people being deceived or unjustifiably impaired in their freedom of choice. Additionally, fairness implies that AI practitioners should respect the principle of proportionality between means and ends, and consider carefully how to “balance competing interests and objectives.31 The procedural dimension of fairness entails the ability to contest and seek effective redress against decisions made by AI systems and by the humans operating them.32 In order to do so, the entity accountable for the decision must be identifiable, and the decision-making processes should be explicable.” (High-Level Expert Group on AI, 2019, p. 13-14) reck5zfiq18a3whae

Specific ethical concepts arising in the context of technology

here to edit

check this, these are now drawn from the same table as the concepts in the introduction and so the thread should be clear.

How is AI research different?

ChallengeInstances Research produces both direct impact, on participants and then society, while also giving rise to indirect, long-range, and dual-use impacts

Application of AI has potential to exacerbate, or ignore, established injustices; these biases may not be immediately clear in unbalanced data, nor apparent in short-term implementation evaluations

"The research analytical process includes selecting the training data, cleaning the data, developing the model through steps of training, evaluating, adjusting, re-training the model.[..] The use of AI systems to uncover or predict social phenomena can be tainted by biases in the training data set on certain demographics or proxies thereof, whic...

ChallengeInstances Privacy, confidentiality, data integrity and cultural norms

In seeking to protect participants we may assume deidentification, however this may serve to marginalise participant contributions to research

Participant we use participant to refer to those choosing to participate in research, those where consent-waivers may be in place or where some stakeholders may fulfil participant-researcher roles (e.g., teachers), and those 'data subjects' whose data is used in research often without their knowledge. input into research should be appropriately protected, and acknowledged, and these may at times be in tension (or, perceived to be in tension where our default assumption is towards deidentification). These issues may become salient in new ways as generative AI emerges, and questions are addressed regarding protection of intellectual...

AI education

ChallengeInstances Interactions and distance between researchers and participants are central to research, both when proximal, and when mediated by technologies.

Technology-mediated relationships between researchers and participants may serve to entrench, create, or miss opportunity to challenge, injustices.

Technology-mediated relationships between researchers and participants may serve to entrench, create, or miss opportunity to challenge, injustices. That is through, for example:- power relationships arising from researcher-directed data selection and use,- lack of certainty regarding who data represents and its underlying meaning, and - potentia...

ChallengeInstances Respect for human-human relationships, at individual and collective levels

Should affective AI nudge users for personal or societal benefit?

"Governmental entities may potentially use nudging strategies, for example to promote the performance of charitable acts. Does the practice of nudging for the benefitof society, including nudgesby affective systems, raiseethical concerns?## BackgroundA few scholars have noted a potentially controversial practice of the future: allowing a robot o...

ChallengeInstances New technologies give rise to new potential for misuse

The complexity of actionable 'accuracy' and reliability in the context of uncertain uses and users: False positives, the tyranny of averages, and operationalising measures

“Because of their statistical basis, all ML systems have error rates. Even though in many cases ML systems are far more accurate than human beings, there is danger in assuming that simply because a system’s predictions are more accurate than a human’s, the outcome is necessarily better. Even if the error rate is close to zero, in a tool with mil...

ChallengeInstances Respect for diversity, non-discrimination, and fairness

Design fairness and ensuring models address challenges communities agree should be addressed

“Because human beings have a hand in all stages of the construction of AI systems, fairness-aware design must take precautions across the AI project workflow to prevent bias from having a discriminatory influence. This requires cross-stakeholder to understand the desired outcomes and appropraite measures, alongside consideration of any biases th...

ChallengeInstances Effective use of AI inherently involves multiple stakeholder and disciplinary expertise, including both AI literacy and contextual awareness across cultures

Assessing the impact of AI requires sensitivity to culturally situated human social interaction

"It is presently unknown** **whether long-term interaction with affective artifacts that lack cultural sensitivity could alter human social interaction.## BackgroundSystems that do not have cultural knowledge incorporated into their knowledge base may or may not interact effectively with humans for whom emotion and culture are significant. Given...

ChallengeInstances Effective use of AI inherently involves multiple stakeholder and disciplinary expertise, including both AI literacy and contextual awareness across cultures

How do we represent cross-cultural differences in communication through AI systems?

"Should affective systems interact using the norms for verbal and nonverbal communication consistent with the norms of the society in which they are embedded?## BackgroundIndividuals around the world express intentions differently, including the ways that they make eye contact, use gestures, or interpret silence. These particularities are part o...

AI affective-computing

And for education

here to edit

check this, these are now drawn from the same table as the concepts in the introduction and so the thread should be clear

How is education research different?

ChallengeInstances Privacy, confidentiality, data integrity and cultural norms

In seeking to protect participants we may assume deidentification, however this may serve to marginalise participant contributions to research

AI education

ChallengeInstances Interactions and distance between researchers and participants are central to research, both when proximal, and when mediated by technologies.

While shifts in researcher-participant relationships can create distance, they can also create connections. However, such research involves power relations that are complex to navigate.

In research that is \[close to practice]\(/recb4ql0nq1tsae0w) has significant potential to directly impact the environment in which research is conducted. However it is important to consider the resource costs of research involvement and who has shaped, will be represented in, and impacted by the research.

education

ChallengeInstances Effective use of AI inherently involves multiple stakeholder and disciplinary expertise, including both AI literacy and contextual awareness across cultures

The Intersection of Data Literacy and Social Justice in AI and decision making

"With growing awareness of, and attention to, the potential of data to inform decisions across contexts, has come an increasing recognition and need to develop data literacy strategies that support people to learn to be critical of data, given this consequential nature of data use (and abuse). To achieve a just society, inequities in both capaci...

education

ChallengeInstances Research produces both direct impact, on participants and then society, while also giving rise to indirect, long-range, and dual-use impacts

Inequities in access to AI may increase, not tackle, inequality

AI tools may be used as a panacea to address well established inequities in resourcing, leading to two-tiered systems including in education; conversely access to AI may be limited by the resources available at a local, regional, or national level in ways that limit further access to effective AI tools and the economic benefits associated with t...

education