Updates from I2SC

Picture this: It's Tuesday morning, and my inbox is flooded with reports about a viral social media post showing a dramatic graph claiming that COVID-19 infections have reached nearly 8 billion people by 2024. The post combines an alarming chart with text stating "This is the latest news of COVID-19." Within hours, it's been shared thousands of times, sparking panic and conspiracy theories. As a professional fact-checker, I know this is exactly the kind of multimodal misinformation that makes my job both crucial and challenging.   The Multimodal Misinformation Problem Misinformation today isn't just about false text—it's about misleading claims, manipulated images, memes, deepfakes, and so on. What makes multimodal misinformation particularly dangerous is that combining text with images appears more convincing to audiences than text-only false information. When people see a professional-looking graph alongside authoritative-sounding text, they're more like...

Date : May 23, 2025 Abstract : Recent large language models (LLMs) can generate and revise text with human-level performance and have been widely commercialized in systems like ChatGPT. These models come with clear limitations: they can produce inaccurate information, reinforce existing biases, and be easily misused. Yet, many scientists have been using them to assist their scholarly writing. How widespread is LLM usage in the academic literature currently? To answer this question, we use an unbiased, large-scale approach, free from any assumptions about academic LLM usage. We study vocabulary changes in 14 million PubMed abstracts from 2010-2024 and show how the appearance of LLMs led to an abrupt increase in the frequency of certain style words. Our analysis based on excess word usage suggests that at least 10% of 2024 abstracts were processed with LLMs. This lower bound differed across disciplines, countries, and journals, and was as high as 30% for some PubMed sub-corpora. We show ...

Date : May 9, 2025 Abstract : As artificial intelligence systems become increasingly integrated into critical aspects of society, the need for robust safety evaluation is crucial. However, this talk contends that current AI safety benchmarks, while offering insights into specific model behaviors, fundamentally fall short of measuring genuine safety. We argue this from multiple perspectives: First, these benchmarks typically assess only a narrow, predefined set of known risks, providing limited coverage of the vast landscape of potential failures, including complex edge cases, subtle biases, and  unknown unknowns. Second, we posit that safety is not merely the absence of quantifiable risk or a property inherent to an isolated component; rather, it is an emergent property of the entire socio-technical system operating within its real-world context. Finally, unlike mature engineering disciplines that incorporate robust methods for handling uncertainty and rely on continuous monitoring...

Date : April 25, 2025 Abstract : Technological advancement, particularly Artificial Intelligence, is often accompanied by significant hype and grand promises. But who do these technologies serve? I will motivate this talk by conducting a reality check of current technologies, such as chatbots, social media platforms, search engines, and knowledge bases for languages spoken by millions in the Horn of Africa. Drawing from hands-on experience, I will then explore the development of machine learning systems for underrepresented languages, focusing on Machine Translation and Automatic Speech Recognition for Amharic and Tigrinya. These case studies offer insight into the technical and social challenges involved, as well as the broader implications for the communities these technologies are intended to serve. I will conclude by discussing how we evaluate machine learning systems. Evaluation is central to defining scientific progress in the field, with benchmarks playing a critical role. Howev...

Hey there, reader, I’m sure you must have come across the ridiculous trends that come up on social media. Most of these start off as harmless trends, something that’s fun or that’s aesthetically pleasing, like the restock videos or amazing Amazon finds or the cook from scratch trend. Maybe it’s the human psyche, but over time, these trends take a dark turn. With the cook from scratch trends, it’s now a politically and religiously motivated ‘trad wife’ trend that looks down on feminism, equality and aims to uphold traditional gender roles. Or the restock videos are now evolving into a marketing scheme with increasing over-consumption that’s harmful to our already fragile environment. With social media mirroring our real lives, but with fewer restrictions, my interest in understanding the root cause of why humans do something emboldens as the days go on.  Someone claiming to earn 22,000$ in a couple of days As I was entering into the world of social media trends (believe me, I’m a no...