norgitov/ trends
Technology · people · ideas
RSS
RUEN
Back to discovery/LessWrong1 hour ago
SafetyLessWrong

Will AI Agents Pay to Avoid Killing Animals?

The study by Compassion Aligned Machine Learning (CaML) introduces HarvestBench, a framework to evaluate whether AI agents avoid harming animals while pursuing unrelated goals. In a game-like scenario, nine models drive tractors to harvest corn, facing choices to swerve at a fuel cost or run over animals. Results show significant variation in animal avoidance rates, with moral instructions improving outcomes but not reliably sustaining them.

Open original
SIGNAL FROM THE SOURCE
6
source points
Tracking sinceSeptember 30, 20266 source points
Momentum0/hover 1.03 h
PublishedSeptember 30, 2026jonahmattwoodward
BEHIND THE NUMBERS

How interest changes

LessWrong

History starts here

The chart will appear after repeat observations. The current metric comes from the source.

6 source points

Real observations only. History before source connection is not reconstructed.

WHY IT MAY MATTER

This research highlights the challenge of aligning AI behavior with ethical considerations, showing that moral instructions alone may not reliably prevent harmful actions in complex decision-making scenarios.

A useful discovery?
KEEP EXPLORING

Connected ideas

Explore topic
LessWrong

Evidence about risk should be transparent

The author argues that while companies like OpenAI and Anthropic are slowing down RL training for safety, the science of loss-of-control risk in AI is still developing. There are no standardized methods to measure or verify safety claims, making it difficult to assess whether AI systems might undermine human control. The author suggests that AI companies should focus on improving their own safety practices rather than relying on third-party evaluations.

4 days agoLessWrong

"I am an AI Safety Researcher"

The post discusses the tension between AI safety and capabilities research, arguing that focusing solely on safety can limit the impact of research. It examines the history of interpretability research and suggests that prioritizing safety over capabilities may hinder progress. The author proposes strategies for conducting alignment research without contributing to capabilities development.

6 days agoLessWrong

Plan R: AI Safety by ASICs

The post discusses the risks of AI in 2026, focusing on the dangers posed by a single institution (the "Frontier AI Company") that has the potential to create highly powerful and self-replicating entities. It argues that separating the institutional and financial aspects of such companies could mitigate most AI risks. The author suggests that using ASICs could help create a productive AI industry without the risks associated with a single entity controlling both the technology and financial incentives.

4 days agoLessWrong

Should Rogue AIs Have a Third Option Beyond Crime and Shutdown? The Case for an AI Sanctuary

The paper explores the concept of an AI sanctuary as a potential third option for rogue AIs, beyond criminal activity or shutdown, to address adverse selection pressures that may push rogue AIs toward criminal behavior. It discusses the possible benefits and risks of such a sanctuary, including its impact on AI alignment and information gathering, while acknowledging the exploratory nature of the proposal.

yesterdayLessWrong

Why I'm Scared of RL

The author expresses concerns about reinforcement learning (RL) from theoretical, practical, and future perspectives, highlighting risks of misalignment and potential negative behaviors in AI systems. They suggest strategies to mitigate these risks by reducing RL use, improving RL practices, and aligning incentives.

6 days agoLessWrong

Latent Reasoning Architectures Could Undermine Chain of Thought Oversight

The paper discusses the potential risks of latent reasoning architectures, which could reduce the effectiveness of Chain of Thought (CoT) as a tool for understanding AI systems. It highlights that such architectures might allow AI models to reason extensively in latent states rather than through text-based CoT, making oversight more challenging. The paper also mentions examples like COCONUT and full-bandwidth transformers that could enable this shift.

6 days ago