AI quality & safety
go hand-in-hand

Papermoon Keeps your AI productS on track

AI empowers your business to grow faster, serve more customer needs, and increase your revenue - but only if it is high quality, trustworthy, and safe. You’re the expert in what your business requires. We’re the experts in AI assessment & risk mitigation.

We help you set the guardrails and monitor your AI, so you can focus on what your customers need.

Need help assessing your risks and quality concerns?
Contact us for a call around services & tools that fit your needs.

Contact Papermoon

Tools

Our tooling helps bridge the gap between policy and implementation teams. We're currently in alpha mode.

Compliance teams that want a bespoke test-drive are welcome to see how we can help save time with first-pass policy review of both human-made and AI-generated content.
‍
Want to know more about what we're building?

Contact Papermoon

Rather want with your larger organization's AI Quality & Safety questions? Check out our services below.

Services

AI risk scan

Establish robust governance frameworks to ensure your AI deployments align with organizational needs and regulatory requirements.

  • Capacity building. Comprehensive introduction to AI Trust & Safety principles tailored to your organization's unique challenges and potential vulnerabilities. Identify the tools and teams you need.
  • Policy & process development. Identify which risks should be a priority, and establish a mitigation roadmap that works in practice. Need more hands-on support? We offer retainers if necessary.  

AI quality tracking

Making an AI prototype is easy, but improving real-world quality requires clear metrics and evaluation.

  • Metrics frameworks. Help your organization establish clear KPIs and clarity on quality metrics that align technical performance with organizational objectives.
  • Audit or evaluate systems. Curate and measure the data you'll need to make informed decisions about existing or new systems. Identify early concerns and opportunities.

Research & investigations

Unexpected behavior can show up after deployment. We help investigate, and deliver deep dives, to assess the situation.

  • User behavior intelligence. Comprehensive examination of behavioral patterns, and anomalous activities, including detailed failure analysis and/or identification of potential bad actors.
  • Algorithmic forensics. Specialized diagnostic services for situations where symptoms are apparent, but root causes remain unclear, requiring in-depth technical analysis.

Team

We are technology veterans with prior leadership roles at research institutes, startups and larger companies like Spotify and Yahoo. We have deep expertise in human-computer interaction, search and recommender systems, language modeling, and AI Trust and Safety. We have built systems from 0 to production, wrangled data instrumentation at scale, investigated crisis escalations, and set up algorithmic responsibility programs.

We're based in San Francisco, California.