Tech giant IBM’s latest offering, watsonx.governance, is touted to revolutionize how businesses and governments use generative AI, hoping to usher transparency and trust in AI (artificial intelligence) models.
Scheduled for a general release in early December, this IBM watsonx.governance aims to simplify the inner workings of AI models by shedding light on both the input data and the resulting outputs.
Generative AI, which took center stage after OpenAI launched ChatGPT last year, driven by Large Language Models (LLMs) or Foundation Models, holds immense potential for various business applications. However, it introduces new challenges such as sourcing training data from unverifiable online sources and producing outputs lacking clear explanations.
IBM unveils $500-million enterprise AI venture fund
IBM expands access to AI education to 2 million learners
Addressing these concerns, watsonx.governance equips organizations with the necessary tools to navigate risks, foster transparency, and proactively align with forthcoming AI-related regulations.
According to IBM, businesses leverage a blend of LLMs from tech providers and open-source communities, watsonx empowers them to oversee, regulate, and govern models regardless of their origin.
Responsible AI
watsonx.governance represents one of the key software products within the IBM watsonx AI and data platform. This platform comprises the watsonx.ai enterprise studio for AI creators and the watsonx.data store, designed to be open, hybrid, and governed. IBM also recently introduced intellectual property protection for its internally developed watsonx models, further emphasizing its commitment to innovation and safeguarding proprietary technology.
IBM Consulting has expanded its expertise to assist clients in responsibly scaling AI. This includes automated model governance and broader organizational governance encompassing people, processes, and technology. Their consultancy spans establishing AI ethics boards, fostering an accountable organizational culture, providing training, managing regulatory compliance and risks, and fortifying against cybersecurity threats, all rooted in human-centric design principles.