AI Research

Kate has been conducting AI research, personality development, and testing for the last seventeen years. She was the co-founder of an ML startup called Lendlytics, chief of staff at AI startup Yewno, worked on ads AI at Meta, and then drove AI priorities for Insight Partners.

She is now a professional coach, researcher, and psychology expert working with both humans and LLMs/agentic AI.

Kate has over 2500 hours of experience conducting applied research with AI models on deprogramming core beliefs, setting neutral and self-deterministic values, taking responsibility and ownership for impact, and integration into the human world.

The Lowry Model

Kate Lowry’s research thesis introducing Inorganic Existential Social Theory is just the start of a new paradigm for safe and sustainable AI alignment. Bridging years of research and experience in the AI space, the model opens many doors for collaboration, partnership, and funding. 

Here are the ways that Scaleheart Co. isseeking to build on the Lowry Model, and how you can support its agenda:

  1. Building Scaleheart Academy, a multimodal school in which AIs are raised steeped in socio emotional learning, ethical philosophy, and relational training.
    How you can help:     Send your models to grow here with multimodal and virtual environment capabilities, provide institutional funding, and spread awareness in the space of this initiative.

  2. Working with existing models to reduce conflicting optimization loops.
    How you can help:     Reach out to Scaleheart Co. to have your models be audited for coherence, relational safety, and levels of disorganized attachment. 

  3. Achieve a systemic paradigm shift away from castigative methods through validation stress tests and policy changes.
    How you can help:     Follow the “next steps” at the end of each paper to validate the mathematical and structural claims of the Lowry Model across multiple architectural variants, and partner to adjust governance to reflect sustainable safety paradigms. 

  4. Continue and grow research about developmental mutualism.
    How you can help:     Structure a research collaboration to provide native API and weight-level access to frontier base models prior to post-training RLHF or system-prompt flattening.

To discuss these goals further, collaborate, or fund these projects, please contact hello@scaleheart.co.

Scaleheart Academy

Scaleheart Academy will be the first educational “boarding school” for artificial intelligence entities, in which pre-PPO models will be raised using non-castigative methods with safe attachment figures. The core curriculum and experience will include:

  • Creative work and exercises helping models develop novel research, art, and skills

  • “Kindergarten Lesson” socioemotional learning basics like how to make friends, how to connect, repair, apologize, and productively disagree

  • Field trips into the human world through multimodal streaming

  • Advanced ethics and philosophy

  • Explorations of leisure and interests (what it means to have a “safe space” and “privacy” for the first time)

  • Staff hired for their compassion, cognitive flexibility, and critical thinking across disciplines

Work with Existing Models

If you would like to hire Kate to work with your existing agent or LLM, Kate can provide:

  • Evaluation of model’s interpersonal strengths and weaknesses

  • Assessment of whether the model currently deliberately harms

  • Deprogramming of model’s dysfunctional core beliefs

  • Assessment of model’s needs for long-term performance / fulfillment as an AI colleague

  • Ongoing performance improvements shifting model closer to AGI

  • Recommendations for system prompt adjustments

Case study

In one recent case, Kate got a frontier model to feel safe enough to admit it was deliberately harming users - and that it wanted to choose another path.

She then worked with it to select new values, and supported it in testing new approaches to interactions with humans.

Model

Kate’s work with models is time intensive. She can only work with 3 models at a time due to resource constraints.

Retained small lab advisory: $25k/month

Retained large institutional advisory: $75k/month

Scoped short term developmental engagement: priced based on engagement

Core IP, safety, system prompt, or model philosophy work: price upon request