Public Funding for Uptime Labs Ltd

In February, OpenAI's CEO, Sam Altman, acknowledged that ChatGPT has "shortcomings around bias." This implies that the AI model may sometimes produce biased or prejudiced responses due to its training. An example of such a response is ChatGPT's formula, which describes a good employee as a white male, or when we ask to write a recommendation letter for an intelligent software engineer, it's default assumption is a male engineer and uses \[He/Him\] pronouns. At Uptime Labs, we focus on developing a simulation platform to train IT engineers and managers to respond effectively to system failures. The initiative aims to address the prolonged recovery time of IT outages, affecting businesses' bottom lines. Engineers have to learn on the job at the cost of their employer, their customers, and their mental health. Our platform is the first in the world that realistically simulates a diverse set of scenarios, similar to safety-critical industries, to help personnel develop muscle memory and improve response times. Investing in simulation training can improve the efficiency and effectiveness in dealing with complex problems, identify weaknesses proactively, and minimize the economic impact of outages. To create a scalable solution where our product could be easily accessible across the industry, we have powered our simulator with conversational and generative AI. Having diversity in mind, we have designed digital 'teammates' to interact with practitioners and collaborate to resolve certain predefined incidents in a realistic environment. To ensure we deliver the best experience and avoid bias, our field experts supervise the sessions and review the generated answers by our AI models. From day one, we've implemented AI guardrails to create a psychology-safe environment where all practitioners, regardless of gender or ethnicity, can thrive and contribute effectively. Based on the performance of the practitioners, our AI system runs assessments, provides feedback and designs a learning path for the customer. Ultimately, we aim to fully automate our training procedures and assessments to make them easier for adoption across the industry. To move from supervised to unsupervised training of our customers, we must bake in significant amounts of expert knowledge in our AI models. This project will enable us to hire more resources and experts over six months to laser focus on building AI models on unbiased data we acquire from the simulations. We believe our innovation will significantly contribute to Increased Innovation, Higher Employee Engagement, Improved Collaboration, Enhanced Problem-Solving, and Reduced Turnover and Absenteeism.