Transdisciplinary AI Observatory—Retrospective Analyses and Future-Oriented Contradistinctions

Abstract

In the last years, artificial intelligence (AI) safety gained international recognition in the light of heterogeneous safety-critical and ethical issues that risk overshadowing the broad beneficial impacts of AI. In this context, the implementation of AI observatory endeavors represents one key research direction. This paper motivates the need for an inherently transdisciplinary AI observatory approach integrating diverse retrospective and counterfactual views. We delineate aims and limitations while providing hands-on-advice utilizing concrete practical examples. Distinguishing between unintentionally and intentionally triggered AI risks with diverse socio-psycho-technological impacts, we exemplify a retrospective descriptive analysis followed by a retrospective counterfactual risk analysis. Building on these AI observatory tools, we present near-term transdisciplinary guidelines for AI safety. As further contribution, we discuss differentiated and tailored long-term directions through the lens of two disparate modern AI safety paradigms. For simplicity, we refer to these two different paradigms with the terms artificial stupidity (AS) and eternal creativity (EC) respectively. While both AS and EC acknowledge the need for a hybrid cognitive-affective approach to AI safety and overlap with regard to many short-term considerations, they differ fundamentally in the nature of multiple envisaged long-term solution patterns. By compiling relevant underlying contradistinctions, we aim to provide future-oriented incentives for constructive dialectics in practical and theoretical AI safety research.

Non-Consensual Synthetic Intimate Imagery: Prevalence, Attitudes, and Knowledge in 10 Countries

“…Fourth, various stakeholders have worked to help create solutions for victims, primarily focusing on triaging after victimization has occurred. It will be necessary for these helpseeking resources (whether helplines, online forms, or chatbots [25,50]) to be aware of this new type of image-based sexual abuse and up-to-date with the types of redress that victims may seek [4]. Fifth, where tools do have a payment or sign up flow, users could be required as part of the flow to ingest and acknowledge information about the harms and potential consequences of NSII creation and As practitioners and designers think about future considerations for intervention, it may be helpful to break this issue down into relevant behaviors as seen in Figure 7 and consider intervention options (which should be considered complementary, as opposed to solely sufficient) at each point in the process.…”

“…While deepfakes can be used in beneficial ways for accessibility and creativity [19,26], abuse potential has increased in recent years as the technology has advanced in sophistication and availability [12,34,53,80]. Deepfakes can be weaponized and used for malicious purposes, including financial fraud, disinformation dissemination, cyberbullying, and sexual extortion ("sextortion") [4,26].…”

VR, Deepfakes and Epistemic Security

“…Indeed, an often overlooked modality is deepfake text [14] -which could be of relevance in multiple SEA AI attack schemes. In practice, while deepfake technology has already been abused for impersonation and cybercrime [15], [16], sextortion and non-consensual voyeurism [17], [18], disinformation and espionage [19], [20], deepfakes in VR [21] may add depth to existing threat vectors next to offering a novel field of affordances for malicious actors -from synthetic non-consensual VR deepfakes [22], [23] to immersive disinformation schemes [24] that could even be extended to educational or scientific settings [25]. Overall, at first sight, it seems that epistemic security considerations caution us against underestimating present-day AI and VR when it comes to answering the following question: do the use and exploit of specific AI and VR technologies risk to harm our own processes of knowledge creation and reasoning?…”

“…"epistemic anarchy" [54], "postepistemic world" [4] and "post-truth era" [55]. Furthermore, the deepfake threat landscape engendered a mechanism called automated disconcertion [17] -the epistemic confusion that arises merely by the possibility of malicious deepfakes. In the light of the aforesaid, it is easily conceivable that adversaries could exploit the contemporary fragile epistemic ecosystem and instrumentalize automated disconcertion.…”

“…On the other hand, in human-computerinteraction (HCI), so-called design fictions enable "HCI and design researchers to co-create, explore and speculate the future" [60]. To augment threat modelling and increase its graspability in security and safety domains, one can craft design fictions grounded in threat models [17] -which may also be relevant for epistemic security in the context of deepfake attacks. In recent years, the concept of co-creation design fictions [61] has been applied to various domains including AI safety [62], AIVR safety [22] and VR security [63].…”