This paper argues that a range of current AI systems have learned how to deceive humans.
We define deception as the systematic inducement of false beliefs in the pursuit of some …

AI alignment aims to make AI systems behave in line with human intentions and values. As
AI systems grow more capable, the potential large-scale risks associated with misaligned AI …

Reinforcement learning from human feedback (RLHF) is a technique for training AI systems
to align with human goals. RLHF has emerged as the central method used to finetune state …

This work identifies 18 foundational challenges in assuring the alignment and safety of large
language models (LLMs). These challenges are organized into three different categories …

In this short consensus paper, we outline risks from upcoming, advanced AI systems. We
examine large-scale social harms and malicious uses, as well as an irreversible loss of …

Foundation models are first pre-trained on vast unsupervised datasets and then fine-tuned
on labeled data. Reinforcement learning, notably from human feedback (RLHF), can further …

In coming decades, artificial general intelligence (AGI) may surpass human capabilities at
many critical tasks. We argue that, without substantial effort to prevent it, AGIs could learn to …

Constructing AI models that respond to text instructions is challenging, especially for
sequential decision-making tasks. This work introduces an instruction-tuned Video …

With extensive pretrained knowledge and high-level general capabilities, large language
models (LLMs) emerge as a promising avenue to augment reinforcement learning (RL) in …

Research in Fairness, Accountability, Transparency, and Ethics (FATE) 1 has established
many sources and forms of algorithmic harm, in domains as diverse as health care, finance …