IBM Data Responsibility Perspective for Watson Data and AI*

Maintain control of your data and build your competitive advantage. Understand how your data is managed.

IBM Data Responsibility Perspective for Watson Data and AI*

Maintain control of your data and build your competitive advantage. Understand how your data is managed.

You control your data

Your data is your most valuable asset. It can yield unique competitive advantage when coupled with the power of AI. But as data – and the models you build with it – becomes more and more valuable, you need to ensure you have control and choice over how it’s used.

At IBM, we believe your data is yours – and yours alone. Therefore, it’s essential to create a system of best practices around data responsibility that guide the secure and ethical management of data, including IBM Watson services and the data on which Watson is trained. This includes contracts and disclosures that help foster full transparency; a strategy that reflects our compliance with existing legislation and policy; and a framework that protects privacy and prioritizes personal data protection.

*IBM Watson Data and AI services are those products found on the Watson platform: https://www.ibm.com/watson/products-services/

Data ownership

Your data is the foundation of your competitive advantage, which is why IBM will not share unique insights derived from your data without your agreement. You are also not required to relinquish rights to your data in order to have the benefits of Watson services. But across the spectrum of data businesses generate, some models may gain more value if you decide to share with others whose innovations may accrue back to you. When you use Watson services on the IBM Cloud, you will have the ability to combine your data sets with other IBM-owned, licensed, or public data sets to yield even more valuable insights.

Security and encryption

IBM employs industry-leading security procedures to safeguard the data with which Watson interacts. This includes use of data encryption, access control methodologies, and proprietary consent management modules, which allow us to code or move data to restrict access to authorized users and to de-identify and use data in accordance with applicable permissions.

Transparency

IBM makes clear when and for what purposes your data is being applied in the solutions we develop and deploy. This may include the major sources of aggregated data and expertise that inform the insights of AI solutions we develop. Third-party or licensed data will be clearly identified. When you use Watson services to enhance your business processes, the unique insights you uncover are your competitive advantage.

Consent and limited use

IBM agreements are transparent. We will not use your data unless you consent to such use, and then we will limit that use to the specific uses clearly described in the agreement. In addition, IBM will not share unique insights derived from your data without your agreement. IBM will return or destroy client content at the request of a client or at the end of the cloud service term.

Watson's three-layer model

Most AI systems that train on your data use it to build their general machine learning models, allowing all users to benefit from that collective knowledge. This can be helpful – but not when you want to maintain your crucial competitive advantage.

Watson is different. We protect your data through a three-layer AI model comprised of a base, an industry, and a custom layer. The layers are designed to share information from the bottom up – ensuring your custom knowledge is not shared with the foundational base or industry layers.

Custom layer

Use your own data and take advantage of the lower layers to build your custom model

Industry layer

Build upon natural language capabilities with industry-specific context
 

Base layer

Tap into core natural language understanding capabilities
 

Here’s how it works: The base layer provides foundational natural language understanding capabilities, while the industry layer adds a domain-specific lens to your data for additional learnings. The custom model on top is built with your data. This architecture and bottom-up approach ensures you train on and learn from your data – while maintaining IP ownership.