"... we propose applying the #risk categories to specific #ai #scenarios, rather than solely to fields of application, using a #riskassessment #model that integrates the #aia [#eu #aiact] with the risk approach arising from the Intergovernmental Panel on Climate Change (#ipcc) and related literature. This model enables the estimation of the magnitude of AI risk by considering the interaction between (a) risk determinants, (b) individual drivers of determinants, and (c) multiple risk types. We use large language models (#llms) as an example."
The paper discusses the risks posed by #artificialintelligence (#ai) systems, from biased lending algorithms to chatbots that spew violent #hatespeech. The author argues that policymakers have a responsibility to consider broader, longer-term #risks from #aitechnology, such as #systemicrisk and the potential for misuse. While #regulatory proposals like the #eu #aiact and the #whitehouse AI Bill of Rights focus on immediate risks, they do not fully address the need for #algorithmicpreparedness. It proposes a roadmap for algorithmic preparedness, which includes five forward-looking principles to guide the development of regulations that confront the prospect of algorithmic black swans and mitigate the harms they pose to society. This approach is particularly important for general purpose systems like #chatgpt, which can be used for a wide range of applications, including ones that may have unintended consequences. The article emphasizes the need for #governance and #regulation to ensure that #aisystems are developed and used in ways that minimize risk and maximize benefit, and it references the #nist AI #riskmanagement Framework as a potential tool for achieving this goal.
" ... it is ... possible to grasp the fil rouge behind the GDPR, the DSA and the AI Act as they express a common constitutional aspiration and direction."