Monitoring AI Services for Misuse

Javadi, Seyyed Ahmad; Norval, Chris; Cloete, Richard; Singh, Jatinder

doi:10.1145/3461702.3462566

Cited by 15 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reasons for this are largely due to the lack of (comprehensive) regulatory provisions and the lack of corresponding technical standards to which they could refer. Where such systems are used, they are often misused [ 44 , 45 ]. This is based on a general lack of knowledge about the realistic possibilities of this technology.…”

Section: Methodsmentioning

confidence: 99%

Sources of Risk of AI Systems

Steimers

Schneider

2022

IJERPH

View full text Add to dashboard Cite

Artificial intelligence can be used to realise new types of protective devices and assistance systems, so their importance for occupational safety and health is continuously increasing. However, established risk mitigation measures in software development are only partially suitable for applications in AI systems, which only create new sources of risk. Risk management for systems that for systems using AI must therefore be adapted to the new problems. This work objects to contribute hereto by identifying relevant sources of risk for AI systems. For this purpose, the differences between AI systems, especially those based on modern machine learning methods, and classical software were analysed, and the current research fields of trustworthy AI were evaluated. On this basis, a taxonomy could be created that provides an overview of various AI-specific sources of risk. These new sources of risk should be taken into account in the overall risk assessment of a system based on AI technologies, examined for their criticality and managed accordingly at an early stage to prevent a later system failure.

show abstract

Section: Methodsmentioning

confidence: 99%

Sources of Risk of AI Systems

Steimers

Schneider

2022

IJERPH

View full text Add to dashboard Cite

show abstract

“…Recognising the opportunity, providers increasingly offer Artificial Intelligence solutions "as a service" (AIaaS) [28,71]. AIaaS is advertised as a way of enabling developers without machine learning expertise to easily add AI capabilities to their applications.…”

Section: Ai As a Service (Aiaas)mentioning

confidence: 99%

“…Next, we comparatively reviewed the websites and marketing materials of AIaaS providers, elaborating the key characteristics, differences and similarities between the services offered, their naming, and the degree of user interaction and control. This process involved aligning the categorisation with the descriptions and terminology used in other AIaaS work [28,70,71,153].…”

Section: Aiaas Taxonomymentioning

confidence: 99%

“…Therefore, we join the growing number of calls urging researchers and developers to shift to a mindset of careful planning and evaluation in order to determine whether an attempt to build an AI model-or in this case an AI service-is appropriate in the first place [15,103,125,133]. While this concern is relevant for any ML model development, it is particularly pertinent in the AIaaS context where potential harms might achieve a far larger scale [28,71]. Before building a new AI service, it is important to consider to what extent (if at all) is it possible for the service to accurately model the social and technical requirements of the various contexts in which it will be deployed.…”

Section: Provider Considerationsmentioning

confidence: 99%

“…In this context, the emergence of the AIaaS paradigm presents a number of challenges. First, given that these services aim to offer sophisticated AI capabilities at low cost and on demand to potentially anyone, often requiring little to no technical expertise, there is a risk that any algorithmic biases or other undesirable behaviours in the AIaaS system could be reproduced on a far greater scale across organisations [28,71]. Further, with AIaaS products developed and sold by private for-profit companies, the inner workings of these commercial algorithms are often opaque or hidden (for various reasons), making it difficult for potential users to determine if, when and why bias manifests itself within these services.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Out of Context: Investigating the Bias and Fairness Concerns of "Artificial Intelligence as a Service"

Lewicki,

Lee,

Cobbe

et al. 2023

Preprint

View full text Add to dashboard Cite

AI as a Service" (AIaaS) is a rapidly growing market, offering various plug-and-play AI services and tools. AIaaS enables its customers (users)-who may lack the expertise, data, and/or resources to develop their own systems-to easily build and integrate AI capabilities into their applications. Yet, it is known that AI systems can encapsulate biases and inequalities that can have societal impact. This paper argues that the context-sensitive nature of fairness is often incompatible with AIaaS' "one-size-fits-all" approach, leading to issues and tensions. Specifically, we review and systematise the AIaaS space by proposing a taxonomy of AI services based on the levels of autonomy afforded to the user. We then critically examine the different categories of AIaaS, outlining how these services can lead to biases or be otherwise harmful in the context of end-user applications. In doing so, we seek to draw research attention to the challenges of this emerging area. CCS CONCEPTS• Computing methodologies → Artificial intelligence; • Social and professional topics → Computing / technology policy; Socio-technical systems.

show abstract