Related papers: Comment on "Learning to Build the Bomb"
This paper investigates the propagation of harmful information in multilingual large language models (LLMs) and evaluates the efficacy of various unlearning methods. We demonstrate that fake information, regardless of the language it is in,…
Information cocoons are frequently cited in the literature on whether and how social media might lead to ideological segregation and political polarization. From the behavioural and communication perspectives, this paper first examines why…
The problem of how to effectively mitigate the flow of misinformation remains a significant challenge. The classical approach to this is public disapproval of claims or "debunking." The approach is still widely used on social media, but it…
This paper investigates the emerging challenges posed by the integration of Artificial Intelligence (AI) in the military domain, particularly within the context of United Nations Security Council Resolution 1540 (UNSCR 1540), which seeks to…
Through systematic empirical investigation, we uncover a fundamental and concerning property of Large Language Models: while they can safely learn facts that don't contradict their knowledge, attempting to update facts with contradictory…
Nuclear energy technologies have the potential to help mitigate climate change. However, these technologies face many challenges, including high costs, societal concern and opposition, and health, safety, environmental and proliferation…
Artificial intelligence offers much promise, but its use in scientific research should be restrained so that the primary aim of academia -- advancing knowledge for humans -- is safeguarded.
We consider the problem of reinforcement learning under safety requirements, in which an agent is trained to complete a given task, typically formalized as the maximization of a reward signal over time, while concurrently avoiding…
We argue that most advances in science and technology during the past and the present centuries rely on modern physics concepts developed in the early $20$th century. The reliance is so profound that any improvement of literacy of science…
Existing physical theories do not predict every feature of our experience but only certain regularities of that experience. That difference between what could be observed and what can be predicted is one kind of limit on scientific…
In this paper I address the difference between knowledge-inquiry and wisdom-inquiry (concepts introduced by N. Maxwell) in nuclear physics education, specifically in senior-level textbooks for first-degree physics students. Following on…
Machine learning has proven to be a valuable tool to approximate functions in high-dimensional spaces. Unfortunately, analysis of these models to extract the relevant physics is never as easy as applying machine learning to a large dataset…
Frontier AI Safety Policies concentrate on prevention: capability evaluations, deployment gates, and usage constraints, while neglecting the capacity to coordinate responses when prevention fails. We argue this coordination gap is…
This paper presents the concept of an adaptive safe padding that forces Reinforcement Learning (RL) to synthesise optimal control policies while ensuring safety during the learning process. Policies are synthesised to satisfy a goal,…
An important part of law and regulation is demanding explanations for actual and potential failures. We ask questions like: What happened (or might happen) to cause this failure? And why did (or might) it happen? These are disguised…
Adversarial training is an effective method to train deep learning models that are resilient to norm-bounded perturbations, with the cost of nominal performance drop. While adversarial training appears to enhance the robustness and safety…
Spreading information through a network of devices is a core activity for most distributed systems. As such, self-stabilizing algorithms implementing information spreading are one of the key building blocks enabling aggregate computing to…
Some education researchers have claimed that we should not teach the Bohr model of the atom because it inhibits students' ability to learn the true quantum nature of electrons in atoms. Although the evidence for this claim is weak, many…
Safe reinforcement learning (RL) trains a policy to maximize the task reward while satisfying safety constraints. While prior works focus on the performance optimality, we find that the optimal solutions of many safe RL problems are not…
Scientific theories of consciousness should be falsifiable and non-trivial. Recent research has given us formal tools to analyze these requirements of falsifiability and non-triviality for theories of consciousness. Surprisingly, many…