Explore IBM’s groundbreaking AI advancements, including new generative models and platform enhancements, revolutionizing enterprise AI deployment and governance
IBM has unveiled its plans for introducing new generative AI foundation models and enhancements to watsonx, its AI and data platform geared towards aiding enterprises in scaling and expediting the impact of AI. These enhancements encompass a technical preview for watsonx.governance, upcoming generative AI data services for watsonx.data, and the integration of watsonx.ai foundation models into select software and infrastructure products.
IBM’s upcoming generative AI models for watsonx.ai include:
- Granite Series Models: IBM has announced the general availability of the initial models in the watsonx Granite model series. This series consists of various generative AI models tailored to advance the incorporation of generative AI into business applications and workflows. Recognizing that a one-size-fits-all model is not suitable for all business use cases, IBM is developing Granite models in different sizes. These IBM models, based on a decoder-only architecture, are designed to facilitate the scaling of AI in businesses. It’s important to note that the standard contractual intellectual property protections for IBM products will apply to IBM-developed watsonx AI models.
- Third-party Models: IBM is now offering Meta’s Llama 2-chat 70 billion parameter model and the StarCoder LLM for code generation within watsonx.ai on IBM Cloud, both of which are currently available.
IBM has put in place a training process for its foundation models that adheres to principles of trust and transparency. This process encompasses rigorous data collection and concludes with control points that enable responsible model and application deployments, addressing aspects such as governance, risk assessment, privacy concerns, bias mitigation, and compliance.
IBM is also unveiling plans to introduce new capabilities across the watsonx platform:
- Tuning Studio: IBM intends to release the initial version of its Tuning Studio, featuring prompt tuning. This feature provides a cost-effective way for clients to adapt foundation models to their specific downstream tasks using their enterprise data, with an anticipated availability in the third quarter of 2023.
- Synthetic Data Generator: IBM has launched a synthetic data generator, enabling users to create artificial tabular data sets from custom data schemas or internal data sets. This empowers users to extract insights for AI model training with reduced risk, facilitating decision-making and expediting time to market. This capability is already available.
- Generative AI: IBM plans to integrate watsonx.ai generative AI capabilities into watsonx.data, enabling users to discover, enhance, visualize, and refine data for AI through a self-service experience driven by a conversational, natural language interface. A technical preview for this feature is anticipated in the fourth quarter of 2023.
- Vector Database Capability: IBM intends to integrate a vector database capability into watsonx.data to support use cases related to watsonx.ai retrieval augmented generation. A technical preview is expected in the fourth quarter of 2023.
- Model Risk Governance for Generative AI: IBM is launching a technical preview for watsonx.governance. During this preview, clients will have access to capabilities for the automated collection and documentation of foundation model details, as well as model risk governance features. These features allow stakeholders to view relevant metrics within their enterprise-wide AI workflows, ensuring appropriate human engagement at crucial junctures.
IBM is actively collaborating with clients across ASEAN countries to facilitate the responsible scaling of AI. This encompasses various stages, from training and validation to tuning and deployment of both traditional machine learning and new generative AI capabilities, all powered by foundation models. Additionally, IBM is assisting in data repository management and enterprise-wide governance processes to control risk effectively.
Furthermore, the IBM watsonx AI and data platform will be complemented by a suite of AI assistants designed to help clients harness the full potential of AI with their trusted data across key enterprise use cases, including:
- Application Modernization: IBM watsonx Code Assistant products are set to launch later this year. They will employ tailored foundation models to transform code and offer code recommendations for developers. IBM has also announced upcoming AI-assisted code products, such as watsonx Code Assistant for Z and watsonx Code Assistant for Red Hat Ansible Lightspeed, to assist developers in various tasks.
- Customer Care: IBM watsonx Assistant will deliver consistent and intelligent customer service solutions using conversational AI. One example is IBM Support Insights Pro, which is expected to be available later this month, using watsonx Assistant to help clients find insights in their multivendor IT infrastructures.
- HR & Talent: IBM watsonx Orchestrate will assist HR professionals in automating repetitive, high-friction tasks and back-office processes, such as interview scheduling or job posting, through a conversational interface.
IBM also plans to integrate watsonx.ai innovations into its hybrid cloud software and infrastructure products, including:
- Intelligent IT Automation: IT Automation products Instana and AIOps Insights will incorporate Intelligent Remediation, featuring watsonx.ai generative AI foundation models to assist IT Ops practitioners in summarizing incident details and providing prescriptive workflow suggestions for faster solutions.
- Developer Services for Watsonx: To facilitate the deployment of AI capabilities closer to clients’ data on IBM Power for SAP workloads, IBM is expected to add SAP ABAP SDK for watsonx to enable AI inference near data on Power systems and deploy AI algorithms on sensitive data and transactions. This is planned for availability in the first quarter of 2024.
IBM’s visionary strides in the realm of AI are poised to reshape the landscape of enterprise-driven innovation. With the introduction of new generative AI foundation models and the expansion of its watsonx platform, IBM is steadfastly committed to fostering trust, transparency, and responsible AI deployment. Through collaboration with clients across ASEAN countries and the integration of AI assistants across key enterprise use cases, IBM is empowering businesses to harness the transformative potential of AI while maintaining control and governance. As these innovations continue to evolve, IBM’s dedication to advancing AI technology remains resolute, promising a future where AI is not just a tool but a trusted partner in driving real-world solutions for businesses worldwide.