Responsible AI

Circadian Connect can:

• Curate dialogue in a workshop mode between stakeholders on ethical and responsible AI practices using certified Science Communication and Science Policy strategies;

• Design in AI impact assessments and Theory of Change for organizations;

• Facilitate Research on AI's societal implications. 
   

Our first project as a vendor was with Google LLC. The purpose was to curate methodology to collect Social Stereotypes and create Dataset to aid in evaluating AI & Machine Learning (ML) models, and mitigating undesirable societal harms.

Circadian Connect co-designed the methodology used in the study and also collected the data. The methodology that is being published (see below) can be used as a benchmark for any NLP (Natural Language Processing) research and practice globally, worth millions of $$$.

 

The methodology developed is under Google's programme called SPICE (Stereotype Pooling in India with Community Engagement) is a major investment in its AI and ML portfolio worth $1 million in India.  

In India, the project is called BiNDI - Bias Intervention for Natural Language Processing and Data in the Indian context.
It is arguably  first ever for a social scientist-led company to bag an inter-disciplinary research partnership.

We co-authored two journal articles with Google researchers, that were presented at international conferences and are in the process of getting published in reputed journals. Details below.