Policies and systems already in place will assist Sama customers in complying with new European Union regulations such as the Corporate Sustainability Due Diligence Directive (CSDDD), German Supply Chain Act, and Sustainable Finance Disclosure Regulation (SFDR). Sama also offers enterprise-grade SSO integrations, GDPR and CCPA compliance, and ISO and TISAX certifications.
SAN FRANCISCO, June 18, 2024 - Sama, the leader in providing data labeling, supervised fine-tuning and model evaluation solutions for some of the world’s biggest artificial intelligence (AI) models, today announced that it meets the standards for the Corporate Sustainability Due Diligence Directive (CSDDD)—Sama is required to meet these regulations as a supplier to large enterprises within the EU—and can support customers in becoming compliant with the Artificial Intelligence Act.
Based on compliance with the German Supply Chain Act, Article 9 of the Sustainable Finance Disclosure Regulation (SFDR), and existing commitments to international standards like the UN Guiding Principles, Sama’s current policies and practices meet the core requirements of the CSDDD.
After achieving 98.6% compliance on a CSR audit aligned with CSDDD standards for a major automotive supplier, Sama was chosen to be the customer’s benchmark for all other data annotation vendors this year, reflecting both companies’ ongoing commitments to aligning with best practices in corporate due diligence and responsible AI.
“We are no longer in AI’s Wild West era. There is only one right way to develop AI, and that’s the responsible way — not just following regulations yourself, but ensuring that your supply chain does too,” said Wendy Gonzalez, CEO of Sama. “We welcome our customers to see firsthand how Sama operates securely, ethically and responsibly. This latest audit both validates our approach and encourages us to continue refining our policies and processes to meet the ongoing, rigorous demands required to keep pace with AI innovation.”
The audit, similar to other third-party evaluations Sama has invited and undergone in the past, included document reviews, presentations, walks of the office floors and random interviews with employees. The purpose of the audit was to ensure that sustainability and environmental stewardship, human rights, health, safety and business ethics standards were embedded into the company’s policies and practices. CSDDD will assess how companies embed human rights and environmental due diligence policies and practices in a similar way. Following its approval at the end of April, EU member countries have two years to implement compliant legislation of their own.
In addition, the AI Act has been approved and will begin going into effect in June 2024. Companies have 12 months to ensure general purpose AI models are compliant; that period extends to 36 months for some models that are defined by the Act as “high risk”. However, given the rapid development in the field, companies must begin transitioning now, including ensuring that high-risk model training data sets are constructed and processed securely according to EU regulations.
Sama employees only work on a project through secure, in-office connections on controlled-access floors, ensuring that only employees working with a project can access associated production floors. Furthermore, Sama has launched integrations with the three major global cloud storage providers, ensuring Sama does not keep a copy of customer data but instead can seamlessly write an updated file directly back to the customer’s server. Sama also offers enterprise-grade SSO integrations, GDPR and CCPA compliance, and ISO and TISAX certifications. By following these international standards, Sama helps its customers comply with both pieces of EU legislation in this area.
For additional high-risk model testing, Sama recently launched Sama Red Team, specifically designed to help proactively improve GenAI and large language models’ safety and reliability.
We've created a scalable solution specifically designed for high-quality, rapid annotation of medium and long sequences of frames, which are essential for automotive AI models. This solution reduces total costs of ownership on complex automotive data for four of the five top OEMs and their Tier 1 suppliers in the world.
Large language models (LLMs) have emerged as powerful tools capable of generating human-like text, understanding complex queries, and performing a wide range of language-related tasks. Creating them from scratch however, can be costly and time consuming. Supervised fine-tuning has emerged as a way to take existing LLMs and hone them to a specific task or domain faster.
Sama provides ML Professionals and AI team Leads with an indispensable solution for Computer Vision data labeling.
