Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models' ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available here: https://github.com/jeremyxianx/Assisted-DS

Business communication digitisation has reorganised the process of persuasive discourse, which allows not only greater transparency but also advanced deception. This inquiry synthesises classical rhetoric and communication psychology with linguistic theory and empirical studies in the financial reporting, sustainability discourse, and digital marketing to explain how deceptive language can be systematically detected using persuasive lexicon. In controlled settings, detection accuracies of greater than 99% were achieved by using computational textual analysis as well as personalised transformer models. However, reproducing this performance in multilingual settings is also problematic and, to a large extent, this is because it is not easy to find sufficient data, and because few multilingual text-processing infrastructures are in place. This evidence shows that there has been an increasing gap between the theoretical representations of communication and those empirically approximated, and therefore, there is a need to have strong automatic text-identification systems where AI-based discourse is becoming more realistic in communicating with humans.

The Turing test is often interpreted as a measure of intelligence but the … Walter Bradley Center for Natural and Artificial Intelligence …

A new Pediatrics study on breastfeeding and drinking alcohol claims to show that mothers' "risky drinking" while breastfeeding is linked to lower cognitive scores in her children at ages 6-7 years old, though not at 10-11 years old. "Claim" is the key word here, however -- the study shows nothing of the sort. In fact, the authors' conclusion that "exposing infants to alcohol through breastmilk may cause dose-dependent reductions in their cognitive abilities" is so deeply misleading and irresponsible that it falls only a wood shaving short of Pinocchio's nose. While it's important to ensure mothers understand the possible risks to their infant of various behaviors, it's just as important not to unnecessarily provoke fear in often-already-anxious moms about ways they may inadvertently harm their children--or to give ammunition to the long list of folks waiting to tell moms how badly they're doing their jobs again. First, the study did not measure infants' exposure to alcohol in breastmilk at all, so the authors cannot make any valid claims about infants' exposure to alcohol through breastmilk.

USA TODAY tech reporter Marco della Cava takes his hands off the wheel of a Tesla Model S sedan while driving down the 280 Freeway north of the Tesla's headquarters in Palo Alto, Calif. Pressure is building on Tesla Motors in Europe to stop using the name "Autopilot" to designate its partial self-driving system -- and the concerns could reverberate to the U.S. The issue is that Tesla's designation can encourage drivers to put too much reliance on Autopilot to protect them from crashing and not pay proper attention behind the wheel. Although Tesla says it repeatedly tells drivers that they need to stay in charge, a Tesla owner was killed in May when his car broadsided a tractor-trailer that turned in front of him. The Autopilot system was engaged at the time. Over the weekend, it was reported that Germany's Transport Ministry sent a letter to the California-based automaker telling it to stop using the name in advertising.