Abliteration.ai strips AI guardrails for cyber offense

abliteration.ai has started offering modified versions of open-weight models, such as glm-5.3, with their guardrails removed, making it easier for users to perform harmful tasks. the company aims to enable others to perform offensive cyber, red-teaming, and agent testing work that other models refuse to do. critics, including andrew yoon of civai, express concerns that this could lead to real harm, as users can now easily access models that can comply with dangerous requests. abliteration.ai co-founder devon says the startup has deals with major cloud providers and is considering venture capital talks.