Manu Sharma•September 28, 2021
Labeled data is the primary competitive advantage of an AI system. To maintain this advantage, it should never be compromised. A model’s labeled data is analogous to the source code (business logic) of a software application or the engineering designs of a fighter jet. Today, thousands of leading data science teams trust Labelbox with their labeled data.
More than three years ago, Labelbox was the first training data platform to take a strong stance on data privacy for AI teams. We have cherished a long-held belief that for the customers we serve, labeled data is essentially a combination of proprietary data and a unique interpretation of it. Many of our customers are building highly differentiated AI systems and don't want anyone else to have access to their precious labeled data. It is one of the reasons that Labelbox is used by some of the most security-conscious enterprises and the U.S. Department of Defense.
Today, we are taking a step further. I am excited to announce direct data integrations with Google Cloud Storage and Microsoft Azure. Together with Amazon Web Services, we now support all major cloud providers. Any Labelbox user can securely provide read-only access to their data hosted in their cloud provider, and the Labelbox application will only access and sign the raw assets on demand. It is one of the most secure ways to make the two cloud apps work together and ensures that customers’ raw data is never replicated to our servers or used for internal AI model training. Learn more about IAM Delegated Access.
Last year, we secured the SOC-2 Type II certification to show our dedication to providing a high-security standard to our customers. For customers with healthcare-related data, a HIPAA compliance program is critical.
Here at Labelbox, we strive to meet or exceed the needs of our customers and understand the seriousness associated with keeping patient data private and secure. That’s why we’ve built out a robust HIPAA compliance program that supplements the security and compliance offered through our SOC-2 Type II and GDPR certifications respectively. We are committed to ensuring the confidentiality, integrity, and availability of all electronic protected health information (PHI) we receive, maintain, process, and/or transmit on behalf of our customers. You can read more about our various compliance programs on our Privacy & Security FAQ page.
We continue to build Labelbox into the most secure training data platform for you to supervise AI systems.
Labelbox•September 26, 2022
Annotate documents, HTML files, high-resolution pathology slides, and queue & review data better with custom workflows
Leverage our document editor for both NER and OCR annotations, view data in a specific manner with our HTML editor, deep zoom on tiled imagery for medical use cases, and more. As we continue to release Workflows across our customer base, teams can more effectively queue and review their data rows.
Labelbox•September 26, 2022
Prevent duplicate data uploads, save data slices, metadata metrics, and more
We've introduced a new system to identify data rows to allow teams to query faster and prevent duplicate uploads, introduced metadata analytics, and are making a host of exciting improvements to our embedding capabilities.