This is a weekly newsletter based on my RSS feeds. I have an LLM pipeline select the most relevant articles from my feeds (based on the about page of this website) and summarize them.
What will society think about AI consciousness? Lessons from the animal case
This document is a scientific article titled “What will society think about AI consciousness? Lessons from the animal case,” published in Trends in Cognitive Sciences. The article explores how societal attitudes towards AI consciousness might be shaped by parallels with attitudes towards animal consciousness, while also considering the key differences between AI and animals that could lead to divergent views. It advocates for interdisciplinary research to understand the possibility of AI consciousness, public attitudes towards it, and the ethical and political norms required to interact with potentially conscious AI systems. It emphasizes the need to start preparing now for the societal implications of AI consciousness.
Constructing language: a framework for explaining acquisition
This document is a research article about language acquisition. The article proposes a constructivist framework for understanding how children learn language, emphasizing structure-building mechanisms, multimodal input, active learning, and dynamic developmental change. It argues that this framework, incorporating these components, provides a valuable lens for developing unified explanations of language acquisition, and discusses how this framework differs from nativist and empiricist theories. The article also touches on computational models, including large language models, and their applicability to human language acquisition.
The effect of degree of prediction error elicited by retrieval on the reconsolidation of fear memory
This ScienceDirect page is an article preview for a study published in the journal Cognition (October 2025). The article, “The effect of degree of prediction error elicited by retrieval on the reconsolidation of fear memory,” investigates how different manipulations of prediction error (PE) during memory retrieval affect fear memory reconsolidation. The study uses a reinforcement learning model to quantify PE and relates it to skin conductance responses. The findings suggest that the overall degree of PE elicited during retrieval, a combination of size, type, and number, influences the reconsolidation process. The study aims to clarify the role of PE in the retrieval-extinction paradigm to promote clinical translation.
Iterated Learning
This page from the Open Encyclopedia of Cognitive Science defines and discusses “Iterated Learning,” a process where individuals learn from others who learned the same way. It highlights that this process, often studied through transmission chains, is influenced by biases in learning and reproduction, ultimately shaping the systems of knowledge or behavior that emerge. The article explores the history, core concepts, experimental applications across various domains (language, music, social stereotypes), current debates, and broader connections to fields like social learning and cultural evolution.
Supervenience
This is an entry on “Supervenience” from the Open Encyclopedia of Cognitive Science. Supervenience describes a relationship where one phenomenon depends on another, such that duplicating the base states would also duplicate the higher-level states. It was initially used in ethics and later in philosophy of mind, however, while useful for expressing determination claims, supervenience doesn’t fully capture the underlying reasons or mechanisms behind these relationships. It can occur without actual determination and is being superseded by the concept of “grounding.” The article discusses core concepts, controversies, and broader connections to other level-linking notions in philosophy and science.