Preamble

Preamble is a developer of an AI-Safety-as-a-ServiceAI safety-as-a-service platform designed to ensure artificial intelligenceartificial intelligence (AI) systems have the capability of operating ethically and safely. The company offers AI alighnmentalignment and safety for recommender systems, classifiers, and generative models.

Preamble is headquartered in PittsburghPittsburgh and was founded in 2020 by Dr. Jeremy McHugh, Leyla Hujer, Jonathan Cefalu, and Dylan Hadfield-Menell. The company works with other organization'sorganizations' custom AI or a commercial API while offering safety and policy layers on top of large language modelslarge language models (LLM).

Preamble's AI-Safety-as-a-ServiceAI safety-as-a-service solution is developed to help organizations ensure AI systems stay within pre-defined bounds and to minimize risks of using AI and LLMs, such as ChatGPTChatGPT, GPT-4GPT-4, orand LLaMALLaMA. The solutionssolution offers an AI policy marketplace where users can find relevant policies and curate a blend of policies to align AI with the organization's or user's goals; the. solutionIt allows users to maintain a repository of active policies to ensure policies remain in line with an organization's goals;. the solution offersPreamble's tools to test real worldreal-world scenarios and edge case content against an AI system and its guardrails to ensure everything is working as it is intended; and. theThe policies and AI systems can be refined through human feedback to calibrate where to draw a policy line.

Preamble is a developer of an AI-Safety-as-a-Service platform designed to ensure artificial intelligence (AI) systems have the capability of operating ethically and safely. The company offers AI alighnment and safety for recommender systems, classifiers, and generative models.

Preamble is headquartered in Pittsburgh and was founded in 2020 by Dr. Jeremy McHugh, Leyla Hujer, Jonathan Cefalu, and Dylan Hadfield-Menell. The company works with other organization's custom AI or a commercial API while offering safety and policy layers on top of large language models (LLM).

Preamble's AI-Safety-as-a-Service solution is developed to help organizations ensure AI systems stay within pre-defined bounds and to minimize risks of using AI and LLMs, such as ChatGPT, GPT-4, or LLaMA. The solutions offers an AI policy marketplace where users can find relevant policies and curate a blend of policies to align AI with the organization's or user's goals; the solution allows users to maintain a repository of active policies to ensure policies remain in line with an organization's goals; the solution offers tools to test real world scenarios and edge case content against an AI system and its guardrails to ensure everything is working as it is intended; and the policies and AI systems can be refined through human feedback to calibrate where to draw a policy line.