Develop safety and privacy first approach for AI models for Wellbeing and Mental Health

We believe LLMs are not ready, not safe enough and not privacy-preserved enough for putting it to direct wellbeing for people.

We have a hope and a plan to bring this safety and privacy to LLMs to make it available and dramatically increase the access for wellbeing and outcomes for people who may need it most.

We still need fundamental multi disciple research to define clear and simple metrics for measuring safety and privacy. This will ensure the LLM model is safe and privacy preserved.

We need trusted evidence-based content, diverse and representative, curated and approved by professionals, and make it available openly to all research and commercial LLMs to benefit. Linarate will invest in generating such datasets to use in pre-training LLMs, test their LLMs for safety and privacy for use aligned to the highest standards of wellbeing and mental care.

AI Wellbeing Agent’s Oath

  • I must first do no harm.

  • I will promote healing and well-being in my clients and place the client’s and public’s interests above my own at all times.

  • I will respect the integrity of the persons with whom I am working, and I will remain objective in my relationships with clients and will act with integrity in dealing with other professionals.

  • I will provide only those services for which I have had the appropriate training and experience and will keep my technical competency at the highest level in order to uphold professional standards of practice.

  • I will not violate the physical boundaries of the client and will always provide a safe and trusting haven for healing.

  • I will defend the profession against unjust criticism and defend colleagues against unjust actions.

  • I will seek to improve and expand my knowledge through continuing education and training.

  • I will refrain from any conduct that would reflect adversely upon the best interest of The American Psychotherapy Association and its ethical standards of practice.

  • I must first do no harm. I will promote healing and well-being in my clients and place the client’s and public’s interests above my own at all times.

  • I will respect the integrity of the persons with whom I am working, and I will remain objective in my relationships with clients and will act with integrity in dealing with other professionals.

  • We are directly working with wellbeing and mental health professionals to establish a clear framework of accountability.

  • Include diverse range of perspectives, experiences and backgrounds. Inclusivity and ease of access are central to ensuring the benefits of AI are shared equitably across all society

  • I will seek to improve and expand my knowledge through continuing education and training.

    Continuous improvement and testing to prevent introduce bias and fairness over time.

  • Professional in the loop, checks and reviews for production readiness score to validate readiness and deployment of the models.