On the purpose of meaningful human control of AI

Davidović, Jovana

doi:10.3389/fdata.2022.1017677

Cited by 3 publications

(1 citation statement)

References 12 publications

(14 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, humans supposedly improve the performance and safety of an AI system (Brennan-Marquez et al 2019: 746). Second, human oversight promotes values such as legitimacy, accountability, dignity, as well as human autonomy and agency (Brennan-Marquez et al 2019: 746;Davidovic 2023; European Commission 2020: 9; High-Level Expert Group on Artificial Intelligence 2019: 16). According to Article 14(2) AIA, keeping humans involved aims at "preventing or minimising the risks to health, safety or fundamental rights", thus addressing both safety and values.…”

Section: Introductionmentioning

confidence: 99%

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Laux

2023

SSRN Journal

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Laux

2023

SSRN Journal

View full text Add to dashboard Cite

Institutionalised distrust and human oversight of artificial intelligence: towards a democratic design of AI governance under the European Union AI Act

Laux

2023

AI & Soc

View full text Add to dashboard Cite

Human oversight has become a key mechanism for the governance of artificial intelligence (“AI”). Human overseers are supposed to increase the accuracy and safety of AI systems, uphold human values, and build trust in the technology. Empirical research suggests, however, that humans are not reliable in fulfilling their oversight tasks. They may be lacking in competence or be harmfully incentivised. This creates a challenge for human oversight to be effective. In addressing this challenge, this article aims to make three contributions. First, it surveys the emerging laws of oversight, most importantly the European Union’s Artificial Intelligence Act (“AIA”). It will be shown that while the AIA is concerned with the competence of human overseers, it does not provide much guidance on how to achieve effective oversight and leaves oversight obligations for AI developers underdefined. Second, this article presents a novel taxonomy of human oversight roles, differentiated along whether human intervention is constitutive to, or corrective of a decision made or supported by an AI. The taxonomy allows to propose suggestions for improving effectiveness tailored to the type of oversight in question. Third, drawing on scholarship within democratic theory, this article formulates six normative principles which institutionalise distrust in human oversight of AI. The institutionalisation of distrust has historically been practised in democratic governance. Applied for the first time to AI governance, the principles anticipate the fallibility of human overseers and seek to mitigate them at the level of institutional design. They aim to directly increase the trustworthiness of human oversight and to indirectly inspire well-placed trust in AI governance.

show abstract

Technology in Suicide Prevention: Fears and Functionality for Crisis Supporters

Hopkins,

Mazzer,

Rickwood

2024

Human Behavior and Emerging Technologies

View full text Add to dashboard Cite

Background: Crisis supporters at Lifeline Australia consistently engage with distressed and often suicidal help‐seekers. The development of technological methods, such as machine learning (ML), in suicide prevention may complement their support work. Investigating attitudes towards the use of ML in crisis support is an important first step.Aims: The current study is aimed at investigating crisis supporters’ attitudes towards ML in crisis support/suicide prevention, beliefs about the effect of technology on the service and help‐seeking, and concerns/opinions about any future technology implementation.Methods: Two hundred fifty‐five crisis supporters aged 20–84 years were recruited through Lifeline Australia. Participants voluntarily completed an anonymous questionnaire, including measures of attitudes towards technology as well as open‐text options, which provided the data for a thematic analysis.Results: Crisis supporters were neutral to negative on an adapted measure of ML use in crisis support. Less than one‐third held the belief that technology would enhance Lifeline services, and over half of the participants felt help‐seekers would be less likely to contact Lifeline if technology was implemented. Thematic analysis of the open‐text questions revealed loss of human connection and mistrust of algorithms to be the most prominent barriers to future technological adoption by Lifeline crisis supporters.Limitations: Clearly defining terms of ML and technology was difficult to do in this hypothetical context, potentially impacting the attitudes expressed.Conclusions: Any new technology to support crisis supporters needs to be carefully codesigned with the workforce to ensure effective implementation and avoid any potential or perceived negative impacts on help‐seekers.

show abstract

On the purpose of meaningful human control of AI

Cited by 3 publications

References 12 publications

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Institutionalised distrust and human oversight of artificial intelligence: towards a democratic design of AI governance under the European Union AI Act

Technology in Suicide Prevention: Fears and Functionality for Crisis Supporters

Contact Info

Product

Resources

About