23 citations · 31 across the 19 of their papers we have counts for
20 papers
Shutdown Sabotage Propensities in Multi-Agent Systems
Amelie Knecht, Ulysse Schaller, Christopher Summerfield +1
The final safeguard against rogue AI behavior is the human ability to shut systems down. It has been theorized that when an AI is instructed to perform a task, self-preservation ca…
AI systems out-persuade expert humans
Kobi Hackenburg, Caroline Wagner, Luke Hewitt +5
Many societal decisions are settled by contests of persuasion. Conversational AI is a powerful new entrant in these contests, but whether it can out-persuade skilled and highly inc…
Running the Gauntlet: Hard Agentic Tasks
Mykola Vysotskyi, Runqi Lin, Grzegorz Biziel +21
As agentic systems continue to evolve and are widely deployed in real-world scenarios, there is a growing demand to faithfully evaluate their capabilities. However, current benchma…
RealityTest: How People Probe AI Identity and Whether Models Disclose It
Anna Gausen, Sarenne Wallbridge, Bessie O'Dell +2
AI systems are increasingly deployed in conversational settings where users may be uncertain whether they are speaking with a human or an AI. Despite mounting regulatory attention…
Post-training makes large language models less human-like
Marcel Binz, Elif Akata, Abdullah Almaatouq +76
Large language models (LLMs) are increasingly used as surrogates for human participants, but it remains unclear which models best capture human behavior and why. To address this, w…
PRISM-X: Experiments on Personalised Fine-Tuning with Human and Simulated Users
Hannah Rose Kirk, Liu Leqi, Fanzhi Zeng +4
Personalisation is a standard feature of conversational AI systems used by millions; yet, the efficacy of personalisation methods is often evaluated in academic research using simu…
Reason to Play: Behavioral and Brain Alignment Between Frontier LRMs and Human Game Learners
Botos Csaba, Sreejan Kumar, Austin Tudor David Andrews +6
Humans rapidly learn abstract knowledge when encountering novel environments and flexibly deploy this knowledge to guide efficient and intelligent action. Can modern AI systems lea…
Single-turn emergency psychiatric triage across 15 frontier AI chatbots
Veith Weilnhammer, Lennart Luettgau, Christopher Summerfield +5
People increasingly turn to general-purpose AI chatbots for advice about emotional and mental health problems, but the ability of these systems to recognize and appropriately triag…
Artificial intelligence can persuade people to take political actions
Kobi Hackenburg, Luke Hewitt, Caroline Wagner +2
There is substantial concern about the ability of advanced artificial intelligence to influence people's behaviour. A rapidly growing body of research has found that AI can produce…
A Multi-Turn Framework for Evaluating AI Misuse in Fraud and Cybercrime Scenarios
Kimberly T. Mai, Anna Gausen, Magda Dubois +5
AI is increasingly being used to assist fraud and cybercrime. However, it is unclear the extent to which current large language models can provide useful information for complex cr…
Disclosure By Design: Identity Transparency as a Behavioural Property of Conversational AI Models
Anna Gausen, Sarenne Wallbridge, Hannah Rose Kirk +2
As conversational AI systems become more realistic and widely deployed, users are increasingly uncertain about whether they are interacting with a human or an AI system. When AI id…
When Do LLM Preferences Predict Downstream Behavior?
Katarina Slama, Alexandra Souly, Dishank Bansal +3
Preference-driven behavior in LLMs may be a necessary precondition for AI misalignment such as sandbagging: models cannot strategically pursue misaligned goals unless their behavio…
Ask don't tell: Reducing sycophancy in large language models
Magda Dubois, Cozmin Ududec, Christopher Summerfield +1
Sycophancy, the tendency of large language models to favour user-affirming responses over critical engagement, has been identified as an alignment failure, particularly in high-sta…
Neural steering vectors reveal dose and exposure-dependent impacts of human-AI relationships
Hannah Rose Kirk, Henry Davidson, Ed Saunders +4
Humans are increasingly forming parasocial relationships with AI systems, and modern AI shows an increasing tendency to display social and relationship-seeking behaviour. However,…
People readily follow personal advice from AI but it does not improve their well-being
Lennart Luettgau, Vanessa Cheung, Magda Dubois +8
People increasingly seek personal advice from large language models (LLMs), yet whether humans follow their advice, and its consequences for their well-being, remains unknown. In a…
Conversational AI increases political knowledge as effectively as self-directed internet search
Lennart Luettgau, Hannah Rose Kirk, Kobi Hackenburg +6
Conversational AI systems are increasingly being used in place of traditional search engines to help users complete information-seeking tasks. This has raised concerns in the polit…
The Levers of Political Persuasion with Conversational AI
Kobi Hackenburg, Ben M. Tappin, Luke Hewitt +7
There are widespread fears that conversational AI could soon exert unprecedented influence over human beliefs. Here, in three large-scale experiments (N=76,977), we deployed 19 LLM…
Lessons from a Chimp: AI "Scheming" and the Quest for Ape Language
Christopher Summerfield, Lennart Luettgau, Magda Dubois +9
We examine recent research that asks whether current AI systems may be developing a capacity for "scheming" (covertly and strategically pursuing misaligned goals). We compare curre…
How malicious AI swarms can threaten democracy: The fusion of agentic AI and LLMs marks a new frontier in information warfare
Daniel Thilo Schroeder, Meeyoung Cha, Andrea Baronchelli +19
Advances in AI offer the prospect of manipulating beliefs and behaviors on a population-wide level. Large language models and autonomous agents now let influence campaigns reach un…
HiBayES: A Hierarchical Bayesian Modeling Framework for AI Evaluation Statistics
Lennart Luettgau, Harry Coppock, Magda Dubois +2
As Large Language Models (LLMs) and other AI systems evolve, robustly estimating their capabilities from inherently stochastic outputs while systematically quantifying uncertainty…