Issue link: https://resources.genetec.com/i/1535086
Since artificial intelligence (AI) algorithms can process large amounts of data very quickly, AI is becoming increasingly prevalent in physical security solutions. However, manufacturers who leverage this technology must adhere to practices that ensure that the resulting solutions can be trusted. We've devised a set of guiding principles for creating, improving, and maintaining our AI models. Our developers follow these guidelines to make sure we apply the same standards where this technology is being employed, independent of our product lines. These solutions are being tested extensively to reduce bias as much as possible, and humans are always kept in the loop to maintain integrity and control. Our approach to Responsible AI Guide There are 3 guiding principles for Responsible AI at Genetec: Privacy and data governance: We only use datasets that respect local data protection regulations and we treat them with the utmost care, limiting access as needed. Trustworthiness and safety: We minimize bias in our AI models, rigorously test them, and strive to always make them explainable. Humans in the loop: We make sure that AI models do not make critical decisions by themselves. We ensure that humans are always in the loop and have the final say. Best practices to ensure Responsible AI use and compliance Conduct risk assessments Evaluate how automating a specific process may impact critical systems or safety protocols. Identify non-critical applications Implement AI into processes that aren't central to your most critical operations. This can help curb major business disruptions. Prioritize human-centered design Ensure that AI applications always empower humans with the information they need to make the best decisions. Take advantage of privacy analytics Deploy built-in privacy features within AI systems to limit and protect access to sensitive information. Broaden data protection strategies Apply cybersecurity measures and best practices to AI-enabled solutions, including handling regular audits and system updates. Choose trusted vendors Work with vendors who consider biases, data and cybersecurity, and Responsible AI principles.