Artificial intelligence (AI) is a field that is actively being worked on, and we see new applications for it emerge on a daily basis. GitHub Copilot enables AI to write code for you, IBM"s Telum chips empower machine learning (ML) models to conduct high volumes of inferencing for real-time sensitive transactions and detect fraud, and Apple is reportedly working on AI technologies that will detect stress, depression, anxiety, and cognitive decline. In fact, Gartner expects the AI software market to reach $62 billion next year.
That said, AI development is always under heavy scrutiny due to its potential for misuse and bias. In the past, AI models have demonstrated racial bias, deepfakes have been used for malicious purposes, and there is also speculation that the technology may soon put many people out of a job. To tackle these problems, UN Educational, Scientific and Cultural Organization (UNESCO) member states recently signed a document that defines the values and principles needed to ensure the healthy development of AI. Now, Microsoft has introduced new resources and tool on this topic as well.
Microsoft has partnered with Boston Consulting Group (BCG) on a document called "Ten Guidelines for Product Leaders to Implement AI Responsibly". This piece of work provides tangible and actionable guidance enabling technical leaders to assess their AI products" benefits and risks, review its design and document it, and test it to ensure that it meets their criteria. You can grab this whitepaper by signing in to Microsoft"s dedicated portal here.
Microsoft has also launched a Responsible AI (RAI) dashboard that offers more governance and control over AI models and data to developers and data scientists. Its capabilities include error analysis, interpretability, and casual inferencing. The project is now generally available, runs on Azure Machine Learning, and is open-source. You can check it out here.
The Redmond tech giant has emphasized that some of its customers including Philips, Scandinavian Airlines, and Novartis are already putting its tools and resource into use. It hopes that more organizations will leverage from these resources in the quest to develop AI products responsibly.