Follow Datanami:
October 2, 2023

IBM Announces Availability of Watsonx Granite Models, New Protections for IBM-Developed Models


IBM, one of the leading providers of hybrid cloud and AI, and consulting services, announced the general availability of the first models in the watsonx Granite model series, which is a collection of generative AI models designed to help businesses build and scale generative AI. The new models, built on decoder-only architecture, are designed to be custom-built so businesses can use them for their specific needs and scale the models as needed. 

IBM has been heavily invested in the foundation models for years. A key objective of the watsonx AI model was to develop base models that offer the highest level of accuracy and precision. IBM’s internal benchmarking shows that they have been successful in their objective. Not only do the new models deliver outstanding accuracy, they require less infrastructure to deliver the results. 

Over the last century, IBM has ushered in powerful new technologies and earned the trust of businesses around the world. With the new watsonx models, IBM has now enabled businesses to be AI value creators. This means that companies can bring their property data into the IBM base model and build a model that is tailored to their use cases and needs. 

The Granite models will be available in different sizes, as IBM has been quick to realize that a single model will not fit different business applications. Businesses can use the Granite models for various applications including using it to summarize long-form content into short descriptions, deploy insight extraction and classification to determine customer sentiment, and retrieve augmented generation for searching enterprise knowledge bases to generate tailored responses to customer inquiries. 

In addition to the proprietary models, IBM will allow businesses to utilize third-party models including Meta’s Llama 2-chat 70 billion parameter model. 

IBM Publishes Information about its Training Methodology for Granite Models

IBM also announced its commitment to advance the responsible and safe use of IBM-created generative AI and foundation models. The IBM foundation models have been trained on business-related datasets from five domains – code, academic, legal, finance, and internet.

According to IBM, the training methodology has been curated for business purposes and includes data filters for objectionable content. In addition, the training data has been benchmarked against external and internal models to help with responsible deployment. This will also help address privacy concerns, risk assessment, governance, and bias mitigation.

IBM also plans to release watsonx.governance later this year. This will be an AI governance toolkit designed to enable trusted AI workflows. End-to-end lifecycle governance will also be used to manage and mitigate client risk with watsonx AI and data platform. 

IBM Intellectual Protections for AI Models

IBM also announced that the contractual intellectual property protections for IBM would apply to the watsonX AI models. This is similar to the standard intellectual property protection for IBM hardware and software products. The goal behind this contractual protection for AI models is to promote AI innovation and build transparency into IBM-developed AI models. Clients can develop AI applications using their own data trusting the IBM foundation models. 

According to Dinesh Nirmal, Senior Vice President, Products, IBM Software.”When it comes to today’s AI innovation boom, the businesses that are positioned for success are the ones outfitted with AI technologies that demonstrate success at scale and have built-in guardrails and practices that enable their responsible use,” 

Mr. Nirmal further added that “Today’s release of IBM’s Granite model series and commitment to stand behind IBM-developed watsonx models is a testament to IBM’s end-to-end model lifecycle management process in its watsonx AI and data platform that delivers businesses cutting-edge AI outfitted for their unique business needs.”

IBM’s latest announcement reaffirms IBM’s dedication to AI innovation and its responsible use. The latest offerings will add to the range of open-source tools and AI governance software created and made available by IBM. The watsonx is just one part of IBM’s generative AI solutions. Businesses can also benefit from IBM’s specialized generative AI expertise to tune and operationalize models for targeted business applications.

Related Items 

IBM Commits to Train 2M in Artificial Intelligence in 3 Years, with a Focus on Underrepresented Communities

IBM to Make Llama 2 Available Within Its Watsonx AI and Data Platform

IBM Accelerates Enterprise AI for Clients with New Capabilities on IBM Z