Holistic AI, a leading enterprise AI governance platform, has announced the launch of Holistic AI OSL, a new open-source library aimed at helping developers build fair and responsible AI systems. This optimized library provides advanced tools for mitigating bias and improving explainability in AI applications. By offering this resource, Holistic AI seeks to empower teams to create more transparent and trustworthy AI systems, fostering innovation that benefits society. The library is available for free download, without any licensing requirements, and can be accessed via Python.

As AI systems are increasingly used in critical areas like recruitment, healthcare, and credit scoring, ensuring fairness and eliminating bias are essential. These systems must avoid inadvertent discrimination and ensure equal treatment across all demographic groups. While AI technology has improved prediction accuracy, 65% of AI researchers and developers still identify bias as a significant issue, according to recent studies.

Holistic AI OSL addresses this challenge by offering tools to tackle five major technical risks in AI systems, ensuring accountability throughout the development process. The library includes key features such as bias mitigation, explainability, robustness, security, and efficacy. The Bias Mitigation feature introduces over 35 bias metrics across five different machine learning tasks and provides 30 strategies that developers can use to eliminate bias in their AI systems. The Explainability tools reveal how AI models make decisions and predictions, helping to build transparency and trust in AI systems by clarifying their behavior.

The library ensures Robustness by enabling models to perform consistently, even when faced with challenges such as adversarial attacks or variations in input data. Security measures are in place to protect user privacy through anonymization and to defend against risks like attribute inference attacks. Finally, the Efficacy of models is enhanced by maintaining accuracy, fairness, robustness, and security under various conditions, supported by rigorous real-world testing.

Adriano Koshiyama, Co-CEO of Holistic AI, emphasized the importance of these tools in improving AI quality. "Our new library equips organizations with solutions for all AI risks, including explainability, robustness, and bias. It supports measurement, reporting, and mitigation throughout the AI lifecycle, offering one of the most advanced tools for enhancing AI applications today," Koshiyama said. "Our goal is to help AI reach its full potential while minimizing risks and compliance issues."

Holistic AI OSL has already garnered interest from industry leaders. MAPFRE, a global insurer operating in nearly 40 countries, is integrating the library into its AI strategy to improve customer experience and operational processes. César Ortega, an expert data scientist at MAPFRE, praised the library for its depth, noting, "This library is not just about identifying AI risks, but actively addressing them with proven, industry-ready techniques. It’s an essential part of any ethical AI development toolkit."By making this open-source library available, Holistic AI is reinforcing its commitment to helping businesses accelerate AI innovation while maintaining accountability and fairness in their systems.