Generative AI entails a credit–blame asymmetry

Mann, Sebastian Porsdam; Earp, Brian D.; Nyholm, Sven; Danaher, John; Møller, Nikolaj; Bowman‐Smart, Hilary; Hatherley, Joshua; Koplin, Julian; Monika, Plozza,; Rodger, Daniel; Treit, Peter V.; Renard, Gregory; McMillan, John; Savulescu, Julian

doi:10.1038/s42256-023-00653-1

Cited by 23 publications

(7 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generating unethical, fraudulent, toxic, violent, pornographic, or other harmful content is a further predominant concern, again focusing notably on LLMs and text-to-image models 25,33,35,53,58,69,[76][77][78][79][80] . Numerous studies highlight the risks associated with the intentional creation of disinformation 35 , fake news 77 , propaganda 80 , or deepfakes 47 , underscoring their significant threat to the integrity of public discourse and the trust in credible media 53,81 . Additionally, papers explore the potential for generative models to aid in criminal activities 82 , incidents of self-harm 72 , identity theft 35 , or impersonation 83 .…”

Section: Safetymentioning

confidence: 99%

“…This pertains to various fields, ranging from customer services to software engineering or crowdwork platforms 35,57 . While new occupational fields like prompt engineering are created 53,81 , the prevailing worry is that generative AI may exacerbate socioeconomic inequalities and lead to labor displacement 35,80 . Additionally, papers debate potential large-scale worker deskilling induced by generative AI 97 , but also productivity gains contingent upon outsourcing mundane or repetitive tasks to generative AI systems 57,93 .…”

Section: Governance -Regulationmentioning

confidence: 99%

See 1 more Smart Citation

Ethical and methodological challenges in building morally informed AI systems

Hagendorff

Danks

2022

AI Ethics

View full text Add to dashboard Cite

Recent progress in large language models has led to applications that can (at least) simulate possession of full moral agency due to their capacity to report context-sensitive moral assessments in open-domain conversations. However, automating moral decision-making faces several methodological as well as ethical challenges. They arise in the fields of bias mitigation, missing ground truth for moral “correctness”, effects of bounded ethicality in machines, changes in moral norms over time, risks of using morally informed AI systems as actual advice, as well as societal implications an increasing importance of algorithmic moral decision-making would have. This paper comments on all these challenges and provides critical considerations for future research on full artificial moral agency. Importantly, some of the adduced challenges can be met by more careful technology design, but others necessarily require engagement with core problems of meta-ethics.

show abstract

Section: Safetymentioning

confidence: 99%

Section: Governance -Regulationmentioning

confidence: 99%

Ethical and methodological challenges in building morally informed AI systems

Hagendorff

Danks

2022

AI Ethics

View full text Add to dashboard Cite

show abstract

“…Ethical guidelines need to be developed to mitigate these risks, particularly in the context of reliable and ethical outcomes. To address these and other risks related to LLMs, Porsdam Mann et al [ 58 ] promoted transparency and engagement in open discussions that will allow LLM developers to demonstrate their commitment to and practice of responsible and ethical practices.…”

Section: Discussionmentioning

confidence: 99%

Potential of Large Language Models in Health Care: Delphi Study

Denecke,

May,

Rivera Romero

2024

J Med Internet Res

View full text Add to dashboard Cite

Background A large language model (LLM) is a machine learning model inferred from text data that captures subtle patterns of language use in context. Modern LLMs are based on neural network architectures that incorporate transformer methods. They allow the model to relate words together through attention to multiple words in a text sequence. LLMs have been shown to be highly effective for a range of tasks in natural language processing (NLP), including classification and information extraction tasks and generative applications. Objective The aim of this adapted Delphi study was to collect researchers’ opinions on how LLMs might influence health care and on the strengths, weaknesses, opportunities, and threats of LLM use in health care. Methods We invited researchers in the fields of health informatics, nursing informatics, and medical NLP to share their opinions on LLM use in health care. We started the first round with open questions based on our strengths, weaknesses, opportunities, and threats framework. In the second and third round, the participants scored these items. Results The first, second, and third rounds had 28, 23, and 21 participants, respectively. Almost all participants (26/28, 93% in round 1 and 20/21, 95% in round 3) were affiliated with academic institutions. Agreement was reached on 103 items related to use cases, benefits, risks, reliability, adoption aspects, and the future of LLMs in health care. Participants offered several use cases, including supporting clinical tasks, documentation tasks, and medical research and education, and agreed that LLM-based systems will act as health assistants for patient education. The agreed-upon benefits included increased efficiency in data handling and extraction, improved automation of processes, improved quality of health care services and overall health outcomes, provision of personalized care, accelerated diagnosis and treatment processes, and improved interaction between patients and health care professionals. In total, 5 risks to health care in general were identified: cybersecurity breaches, the potential for patient misinformation, ethical concerns, the likelihood of biased decision-making, and the risk associated with inaccurate communication. Overconfidence in LLM-based systems was recognized as a risk to the medical profession. The 6 agreed-upon privacy risks included the use of unregulated cloud services that compromise data security, exposure of sensitive patient data, breaches of confidentiality, fraudulent use of information, vulnerabilities in data storage and communication, and inappropriate access or use of patient data. Conclusions Future research related to LLMs should not only focus on testing their possibilities for NLP-related tasks but also consider the workflows the models could contribute to and the requirements regarding quality, integration, and regulations needed for successful implementation in practice.

show abstract

“…Generative AI systems put pressure on existing ethical concepts and intuitions. For instance, generative AI creates a new credit-blame asymmetry when assigning responsibility for language model outputs, in that human users should still be blamed for utilizing bad or lowquality outputs of those systems, yet should not get (as much) credit for utilizing particularly good outputs [135].…”

Section: Case-study 2: the Rise Of Generative Aimentioning

confidence: 99%

“…Where A related challenge is that, as LLMs begin to change the nature of many workplace tasks, the aforementioned growing credit-blame asymmetry will begin to express itself in an 'achievement gap', whereby many human jobs will involve supervising, prompting or maintaining LLMs to produce the outputs that skilled humans previously received credit for; but where it becomes increasingly hard for human professionals to claim credit for these tasks [135]; this may lead to a reappreciation of the nature and value of meaningful work, which might be taken as a need for regulatory updating in domains such as labor and employment law.…”

Section: Case-study 2: the Rise Of Generative Aimentioning

confidence: 99%

The technology triad: disruptive AI, regulatory gaps and value change

Hopster

Maas²

2023

AI Ethics

View full text Add to dashboard Cite

Disruptive technologies can have far-reaching impacts on society. They may challenge or destabilize cherished ethical values and disrupt legal systems. There is a convergent interest among ethicists and legal scholars in such “second-order disruptions” to norm systems. Thus far, however, ethical and legal approaches to technological norm-disruption have remained largely siloed. In this paper, we propose to integrate the existing ‘dyadic’ models of disruptive change in the ethical and legal spheres, and shift focus to the relations between and mutual shaping of values, technology, and law. We argue that a ‘triadic’ values-technology-regulation model—“the technology triad”—is more descriptively accurate, as it allows a better mapping of second-order impacts of technological changes (on values and norms, through changes in legal systems—or on legal systems, through changes in values and norms). Simultaneously, a triadic model serves to highlight a broader portfolio of ethical, technical, or regulatory interventions that can enable effective ethical triage of—and a more resilient response to—such Socially Disruptive Technologies. We illustrate the application of the triadic framework with two cases, one historical (how the adoption of the GDPR channeled and redirected the evolution of the ethical value of ‘privacy’ when that had been put under pressure by digital markets), and one anticipatory (looking at anticipated disruptions caused by the ongoing wave of generative AI systems).

show abstract

Generative AI entails a credit–blame asymmetry

Cited by 23 publications

References 17 publications

Ethical and methodological challenges in building morally informed AI systems

Ethical and methodological challenges in building morally informed AI systems

Potential of Large Language Models in Health Care: Delphi Study

The technology triad: disruptive AI, regulatory gaps and value change

Contact Info

Product

Resources

About