Power to the People? Opportunities and Challenges for Participatory AI

Birhane, Abeba; Isaac, William; Prabhakaran, Vinodkumar; Díaz, Mark; Elish, Madeleine Clare; Gabriel, Iason; Mohamed, Shakir

doi:10.1145/3551624.3555290

Cited by 102 publications

(55 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Although complex, it seems prudent and timely to tackle the issue of how to encourage populations to participate in the development of AI more broadly [4]. We consolidate our preliminary analysis into four categories of potential focus for the design of participatory AI for creatives: (1) Understanding AI, (2) Coping with AI, (3) Adapting to AI, and (4) Exploiting AI.…”

Section: Discussion: Opportunities For Participationmentioning

confidence: 99%

“…One response to perceived threats posed by AI is the notion of participatory AI, where the goal is to include 'wider publics' in the development and deployment of AI systems [4]. Historically, the emergence of participatory design (PD) in the 1970s was motivated by efforts to rebalance "power and agency in the professional realm" [3] in order to empower workers to "codetermine the development of the information system and of their workplace" [8].…”

mentioning

confidence: 99%

“…Historically, the emergence of participatory design (PD) in the 1970s was motivated by efforts to rebalance "power and agency in the professional realm" [3] in order to empower workers to "codetermine the development of the information system and of their workplace" [8]. In light of this historical backdrop, participatory AI is expected to empower those affected by the development of novel technologies by enforcing values of inclusion, plurality, collective safety, and ownership [4].…”

mentioning

confidence: 99%

See 2 more Smart Citations

Designing Participatory AI: Creative Professionals’ Worries and Expectations about Generative AI

Inie

Falk

Tanimoto

2023

Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

Generative AI, i.e., the group of technologies that automatically generate visual or written content based on text prompts, has undergone a leap in complexity and become widely available within just a few years. Such technologies potentially introduce a massive disruption to creative fields. This paper presents the results of a qualitative survey (𝑁 = 23) investigating how creative professionals think about generative AI. The results show that the advancement of these AI models prompts important reflections on what defines creativity and how creatives imagine using AI to support their workflows. Based on these reflections, we discuss how we might design participatory AI in the domain of creative expertise with the goal of empowering creative professionals in their present and future coexistence with AI. CCS CONCEPTS• Computing methodologies → Philosophical/theoretical foundations of artificial intelligence; • Human-centered computing → Empirical studies in HCI; HCI design and evaluation methods.

show abstract

Section: Discussion: Opportunities For Participationmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Designing Participatory AI: Creative Professionals’ Worries and Expectations about Generative AI

Inie

Falk

Tanimoto

2023

Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…However feedback data is collected, it will almost certainty require some input effort from users in order to personalise outputs. While this process is participatory, it risks being extractivea form of volunteer labour on the part of end-users for the benefit or profit of technology providers [28]. Volunteer labour to shape the internet landscape has analogies in consumers writing product reviews [196] and social media users flagging content [74].…”

Section: Ir1 Effortmentioning

confidence: 99%

Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback

Kirk¹,

Vidgen²,

Röttger³

et al. 2023

Preprint

View full text Add to dashboard Cite

Large language models (LLMs) are used to generate content for an increasingly wide range of tasks, and are set to reach a growing audience in coming years due to integration in product interfaces like ChatGPT or search engines like Bing. This intensifies the need to ensure that models are aligned with human preferences and do not produce unsafe, inaccurate or toxic outputs. While alignment techniques like reinforcement learning with human feedback (RLHF) and red-teaming can mitigate some safety concerns and improve model capabilities, it is unlikely that an aggregate fine-tuning process can adequately represent the full range of users' preferences and values. Different people may legitimately disagree on their preferences for language and conversational norms, as well as on values or ideologies which guide their communication. Personalising LLMs through micro-level preference learning processes may result in models that are better aligned with each user. However, there are several normative challenges in defining the bounds of a societally-acceptable and safe degree of personalisation. In this paper, we ask how, and in what ways, LLMs should be personalised. First, we review literature on current paradigms for aligning LLMs with human feedback, and identify issues including (i) a lack of clarity regarding what alignment means; (ii) a tendency of technology providers to prescribe definitions of inherently subjective preferences and values; and (iii) a "tyranny of the crowdworker", exacerbated by a lack of documentation in who we are really aligning to. Second, we present a taxonomy of benefits and risks associated with personalised LLMs, for individuals and society at large. Finally, we propose a three-tiered policy framework that allows users to experience the benefits of personalised alignment, while restraining unsafe and undesirable LLM-behaviours within (supra-)national and organisational bounds.

show abstract

“…La creación de valor público se basa en una expresión políticamente mediada de preferencias determinadas colectivamente; es decir, lo que la ciudadanía considera valioso (Trischler et al, 2019(Trischler et al, , p. 1596) y lo que las partes interesadas que participan en la implementación consideran valioso, teniendo conjuntamente el potencial de ofrecer resultados o beneficios para la sociedad. Estas extensiones ponen de manifiesto la importancia de considerar la naturaleza multiactoral de las actividades de co-creación de valor (Birhane et al, 2022). Sin embargo, en cuanto a su importancia para la conceptualización del diseño de las diferentes estrategias de ia, la lógica dominante de estos, al ser la base fundamental para consolidar una regulación o unas reglas de juego para los actores, implica que no es posible diseñar estos denominados 'marcos' como productos predefinidos.…”

Section: Pensamiento De Diseño Y La Inclusión De Las Múltiples Partes...unclassified

Pensamiento de diseño y marcos éticos para la Inteligencia Artificial: una mirada a la participación de las múltiples partes interesadas

Rojas

2023

DESAFÍOS

View full text Add to dashboard Cite

Los marcos regulatorios, que determinan la participación de las partes interesadas, deben respetar los principios para una participación eficaz y que sea oportuna, inclusiva, transparente e iterativa. Sin embargo, la creación e implementación de marcos regulatorios sobre temas innovadores, en ocasiones, suele estar sesgada por una parte dominante en la discusión. Este artículo demuestra cómo el enfoque del pensamiento de diseño, principalmente en su fase de co-creación, es potencialmente propicio para la creación e implementación de las denominadas estrategias de Inteligencia Artificial (ia), debido a su énfasis en el respeto por la dignidad humana y otros atributos fundamentales, como la empatía, aunque plantea diversos retos para su aplicación. Este enfoque se basa en una metodología deductiva, a través del proceso de creación y socialización de las estrategias para la ia, que fueron desarrollados en Canadá y Colombia. Con este artículo se busca incentivar la innovación tecnológica y regulatoria a través del pensamiento de diseño para que sea considerado como parte integral de la convocatoria real de las múltiples partes interesadas, con el fin de facilitar la creación de un programa proactivo de ética digital para prevenir las preocupaciones relacionadas con la adopción de la ia a gran escala.

show abstract

Power to the People? Opportunities and Challenges for Participatory AI

Cited by 102 publications

References 35 publications

Designing Participatory AI: Creative Professionals’ Worries and Expectations about Generative AI

Designing Participatory AI: Creative Professionals’ Worries and Expectations about Generative AI

Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback

Pensamiento de diseño y marcos éticos para la Inteligencia Artificial: una mirada a la participación de las múltiples partes interesadas

Contact Info

Product

Resources

About