![](https://www.techmonitor.ai/wp-content/uploads/sites/29/2023/05/shutterstock_1647111268.jpg)
IBM is entering the artificial intelligence platform market with a new service offering – a toolkit called “watsonx”. The company says it has infused its existing Watson product range with open-source foundation AI models from Hugging Face and other vendors.
![IBM says watsonx will have a range of open source foundation AI models available (Photo: Tada Images / Shutterstock)](https://techmonitor.ai/wp-content/uploads/sites/4/2023/05/shutterstock_1647111268.jpg)
Announced at the Think conference, the Big Blue says watsonx has been launched for code, AIOps, digital labour, security, and sustainability. A new Centre of Excellence for generative AI has also been established to help businesses understand the value of enterprise-grade artificial intelligence.
Since the launch of ChatGPT from OpenAI in November 2022 companies have looked to generative AI, foundation models and other forms of artificial intelligence as a transformative technology. Deploying these models comes with data privacy risks, and experts have warned that vendors need to find enterprise-safe solutions.
IBM says its solution offers foundation models and machine learning capabilities for enterprise using trusted data, speed and governance through a secure cloud environment. They also plan to offer GPU-as-a-service infrastructure for more AI-intensive workloads.
Watsonx will include a new AI development studio that has access to a range of foundation AI models “curated by IBM”. They will all have access to a data store to enable gathering and cleansing of training and tuning data, the company explained.
“With the development of foundation models, AI for business is more powerful than ever,” said Arvind Krishna, IBM CEO and chairman. “Foundation models make deploying AI significantly more scalable, affordable, and efficient.”
“We built IBM watsonx for the needs of enterprises, so that clients can be more than just users, they can become AI advantaged. With IBM watsonx, clients can quickly train and deploy custom AI capabilities across their entire business, all while retaining full control of their data.”
IBM watsonx features cleaned and tuned AI models
There are three core product sets within the watsonx range including watsonx.ai which is an enterprise AI studio due to launch in July. It is for builders of AI tools to train, test, tune and deploy machine learning and generative AI systems powered by foundation models.
IBM says its models use a large, curated set of enterprise data that is backed by robust filtering, a cleansing process and auditable data lineage. Data authority and security has proved a sticking point for large enterprises considering moving to foundation models. IBM says its approach allows for scalable, monitorable solutions.
The multi-modal models are trained on code, time-series data, tabular data, IT events data and geospatial data and are designed to work in partnership with one another. Models will be split into a series of categories with those designed to automatically generate code through natural language interfaces, as well as those allowing the utilisation of curated data where bias can be mitigated more easily and can be quickly customized using client data.
“To benefit from the latest AI capabilities, enterprises want to build on open-source machine learning: open-source models trained on accessible datasets that can run within a secure environment with compliance and proper data governance,” said Clem Delangue, co-founder and CEO of Hugging Face. “We are excited to partner with IBM to deliver a top-tier developer experience based on Hugging Face open-source and community-driven machine learning, within the new watsonx platform offering the enterprise-readiness and trustworthiness of IBM.”
There will also be a new watsonx data store that is built on the lakehouse architecture. It has been optimised for data and AI workloads and will launch in July allowing for on-premises and multi-cloud deployment. IBM claims that “through workload optimization, with this solution, an organization can reduce data warehouse costs by up to 50%.”
Governance tools are likely to be among the most useful for enterprise deployment of foundation model AI. It includes a toolkit to enable trusted workflows and mechanisms to operationalise the governance process. This will mitigate risk, time and cost of providing documentation for explainable and transparent outcomes – both requirements of the EU AI Act and UK AI framework.
IBM says foundation models will be included in all of its major software products going forward.