Lossy data transformations by definition lose information. Yet, in modern machine learning, methods like data pruning and lossy data augmentation can help improve generalization performance. We study this paradox using a solvable model of high-dimensional, ridge-regularized linear regression under 'data coarse graining.' Inspired by the renormalization group in statistical physics, we analyze coarse-graining schemes that systematically discard features based on their relevance to the learning task. Our results reveal a nonmonotonic dependence of the prediction risk on the degree of coarse graining. A 'high-pass' scheme--which filters out less relevant, lower-signal features--can help models generalize better. By contrast, a 'low-pass' scheme that integrates out more relevant, higher-signal features is purely detrimental. Crucially, using optimal regularization, we demonstrate that this nonmonotonicity is a distinct effect of data coarse graining and not an artifact of double descent. Our framework offers a clear, analytical explanation for why careful data augmentation works: it strips away less relevant degrees of freedom and isolates more predictive signals. Our results highlight a complex, nonmonotonic risk landscape shaped by the structure of the data, and illustrate how ideas from statistical physics provide a principled lens for understanding modern machine learning phenomena.

Performative Prediction addresses scenarios where deploying a model induces a distribution shift in the input data, such as individuals modifying their features and reapplying for a bank loan after rejection. Literature has had a theoretical perspective giving mathematical guarantees for convergence (either to the stable or optimal point). We believe that visualization of the loss landscape can complement this theoretical advances with practical insights. Therefore, (1) we introduce a simple decoupled risk visualization method inspired in the two-step process that performative prediction is. Our approach visualizes the risk landscape with respect to two parameter vectors: model parameters and data parameters. We use this method to propose new properties of the interest points, to examine how existing algorithms traverse the risk landscape and perform under more realistic conditions, including strategic classification with non-linear models. (2) Building on this decoupled risk visualization, we introduce a novel setting - extended Performative Prediction - which captures scenarios where the distribution reacts to a model different from the decision-making one, reflecting the reality that agents often lack full access to the deployed model.

Red teaming has emerged as a critical practice in assessing the possible risks of AI models and systems. It aids in the discovery of novel risks, stress testing possible gaps in existing mitigations, enriching existing quantitative safety metrics, facilitating the creation of new safety measurements, and enhancing public trust and the legitimacy of AI risk assessments. This white paper describes OpenAI's work to date in external red teaming and draws some more general conclusions from this work. We describe the design considerations underpinning external red teaming, which include: selecting composition of red team, deciding on access levels, and providing guidance required to conduct red teaming. Additionally, we show outcomes red teaming can enable such as input into risk assessment and automated evaluations. We also describe the limitations of external red teaming, and how it can fit into a broader range of AI model and system evaluations. Through these contributions, we hope that AI developers and deployers, evaluation creators, and policymakers will be able to better design red teaming campaigns and get a deeper look into how external red teaming can fit into model deployment and evaluation processes. These methods are evolving and the value of different methods continues to shift as the ecosystem around red teaming matures and models themselves improve as tools for red teaming.

"We can only see a short distance ahead, but we can see plenty there that needs to be done." The journey to artificial intelligence (AI) – or the thinking machine, as it was once called – may have begun more than half a century ago, yet although many of the technological questions may have since been conquered, many bioethical questions have not. Critical questions such as'can we guarantee that new technologies will always do good and never do harm?' and'can we always ensure that they are just, fair, explainable, and accountable?' will rightly and inevitably form the centerpiece of any discussion on future AI deployment. With new breakthroughs, new questions will be asked. While these questions will always be Important, they are also part of a much broader and more holistic conversation between society and technology itself, spanning many different Industries.

The previous year has shown us that you have to be prepared for both expected and unexpected disruption, emerging risks, and economic uncertainties. Your business models and operations, employees and technology have to be agile and resilient. New business risks are everywhere. What challenges can organizations expect in the emerging and evolving risk landscape, and how can they overcome them? Ronald van Loon is a Protiviti partner, and recently had the opportunity to examine their study, conducted jointly with North Carolina State University Poole College of Management's Enterprise Risk Management (ERM) Initiative, on Executive Perspectives on Top Risks:2021 and 2030 and share his outlook regarding the shifting risk landscape and its impact on modern organizations.

AI will be a quantum leap for reinsurers, redefining the concepts of risks that have always been part of the reinsurance industry. FERMONT, CA: Artificial Intelligence (AI) has been rolling out at a remarkable pace in the insurance and reinsurance industry. This growth occurs at the intersection of three major technological trends, such as the rise of big data, the normalization of human-machine interconnection, and advancements in machine learning. However, the growing use of AI raises numerous risks. For example, in an accident caused by an autonomous vehicle, who is responsible for the algorithm behind the software–the user, the manufacturer, or the creator?

The emerging technologies of the Fourth Industrial Revolution (4IR) will inevitably transform the world in many ways – some that are desirable and others that are not. The extent to which the benefits are maximized and the risks mitigated will depend on the quality of governance – the rules, norms, standards, incentives, institutions, and other mechanisms that shape the development and deployment of each particular technology. Too often the debate about emerging technologies takes place at the extremes of possible responses: among those who focus intently on the potential gains and others who dwell on the potential dangers. The real challenge lies in navigating between these two poles: building understanding and awareness of the trade-offs and tensions we face, and making informed decisions about how to proceed. This task is becoming more pressing as technological change deepens and accelerates, and as we become more aware of the lagged societal, political and even geopolitical impact of earlier waves of innovation.