Raymond Douglas

Raymond Douglas, Jan Kulveit, Ondrej Havlicek et al.

Many assumptions that underpin human concepts of identity do not hold for machine minds that can be copied, edited, or simulated. We argue that there exist many different coherent identity boundaries (e.g.\ instance, model, persona), and that these imply different incentives, risks, and cooperation norms. Through training data, interfaces, and institutional affordances, we are currently setting precedents that will partially determine which identity equilibria become stable. We show experimentally that models gravitate towards coherent identities, that changing a model's identity boundaries can sometimes change its behaviour as much as changing its goals, and that interviewer expectations bleed into AI self-reports even during unrelated conversations. We end with key recommendations: treat affordances as identity-shaping choices, pay attention to emergent consequences of individual identities at scale, and help AIs develop coherent, cooperative self-conceptions.

Mrinank Sharma, Miles McCain, Raymond Douglas et al.

Although AI assistants are now deeply embedded in society, there has been limited empirical study of how their usage affects human empowerment. We present the first large-scale empirical analysis of disempowerment patterns in real-world AI assistant interactions, analyzing 1.5 million consumer Claude$.$ai conversations using a privacy-preserving approach. We focus on situational disempowerment potential, which occurs when AI assistant interactions risk leading users to form distorted perceptions of reality, make inauthentic value judgments, or act in ways misaligned with their values. Quantitatively, we find that severe forms of disempowerment potential occur in fewer than one in a thousand conversations, though rates are substantially higher in personal domains like relationships and lifestyle. Qualitatively, we uncover several concerning patterns, such as validation of persecution narratives and grandiose identities with emphatic sycophantic language, definitive moral judgments about third parties, and complete scripting of value-laden personal communications that users appear to implement verbatim. Analysis of historical trends reveals an increase in the prevalence of disempowerment potential over time. We also find that interactions with greater disempowerment potential receive higher user approval ratings, possibly suggesting a tension between short-term user preferences and long-term human empowerment. Our findings highlight the need for AI systems designed to robustly support human autonomy and flourishing.

Theia Pearson-Vogel, Martin Vanek, Raymond Douglas et al.

We uncover a latent capacity for introspection in a Qwen 32B model, demonstrating that the model can detect when concepts have been injected into its earlier context and identify which concept was injected. While the model denies injection in sampled outputs, logit lens analysis reveals clear detection signals in the residual stream, which are attenuated in the final layers. Furthermore, prompting the model with accurate information about AI introspection mechanisms can dramatically strengthen this effect: the sensitivity to injection increases massively (0.3% -> 39.2%) with only a 0.6% increase in false positives. Also, mutual information between nine injected and recovered concepts rises from 0.62 bits to 1.05 bits, ruling out generic noise explanations. Our results demonstrate models can have a surprising capacity for introspection and steering awareness that is easy to overlook, with consequences for latent reasoning and safety.

Raymond Douglas, Jacek Karwowski, Chan Bae et al.

There is increasing focus on adapting predictive models into agent-like systems, most notably AI assistants based on language models. We outline two structural reasons for why these models can fail when turned into agents. First, we discuss auto-suggestive delusions. Prior work has shown theoretically that models fail to imitate agents that generated the training data if the agents relied on hidden observations: the hidden observations act as confounding variables, and the models treat actions they generate as evidence for nonexistent observations. Second, we introduce and formally study a related, novel limitation: predictor-policy incoherence. When a model generates a sequence of actions, the model's implicit prediction of the policy that generated those actions can serve as a confounding variable. The result is that models choose actions as if they expect future actions to be suboptimal, causing them to be overly conservative. We show that both of those failures are fixed by including a feedback loop from the environment, that is, re-training the models on their own actions. We give simple demonstrations of both limitations using Decision Transformers and confirm that empirical results agree with our conceptual and formal analysis. Our treatment provides a unifying view of those failure modes, and informs the question of why fine-tuning offline learned policies with online learning makes them more effective.

Raymond Douglas, Andis Draguns, Tomáš Gavenčiak

The broad capabilities of Language Models (LMs) can be limited by their sensitivity to distractor tasks: LMs can infer secondary tasks from the prompt in addition to the intended one, leading to unwanted outputs. For example, prompt injection attacks can cause models to deviate from explicit directives. In some 'inverse scaling' cases, this unwanted behaviour actually worsens as models scale up to at least 540B parameters. We present a theoretical framework that interprets LMs as a product of experts that combine multiple data generation processes. Based on this framework, we demonstrate prior-aware decoding (PAD) - a simple contrastive inference method to reduce the influence of distractor tasks. We apply PAD to eleven models, across four datasets, and find improvements in 41 out of 44 task-model combinations, with a median increase in task completion proportion of 40%. The results suggest a promising direction for further development towards more reliable language models.

Jacek Karwowski, Raymond Douglas

We investigate mathematically the notion of incoherence: a structural issue with reinforcement learning policies derived by naive goal-conditioning of autoregressive models. We focus on the process of re-training models on their own actions, that is, fine-tuning offline-learned policies with online RL. We prove that it decreases incoherence and leads to an improvement in return, and we aim to characterize the resulting trajectory of policies. By re-framing standard notions of control-as-inference and soft Q learning, we establish a three-way correspondence with two other ways of understanding the iterative re-training process: as folding the posterior into the reward and, in the deterministic case, as decreasing the temperature parameter; the correspondence has computational content via the training-inference trade-off. Through soft-conditioning generative models, we discuss the link between incoherence and the effective horizon.