Fiddler AI, a leading company in the field of AI observability and security, is tirelessly working to make the world a better place. At the recent Generative AI Summit held in Washington D.C., Nick Nolan, Solutions Engineering Manager at Fiddler AI, shared information about the company’s initiatives in the field of artificial intelligence and its commitment to developing responsible applications.
Fiddler AI’s mission is focused on building trust in AI systems, addressing critical areas such as creating responsible AI applications and ensuring that users receive fair and reliable answers when interacting with these technologies. As artificial intelligence advances, especially in content generation, it is expected that new policies will emerge to regulate governance, risk, and compliance, which is crucial to protect consumers from harmful or biased outcomes.
To guide companies in the responsible use of artificial intelligence, Fiddler AI is developing its approach in three key areas: monitoring, security, and governance. Their observability platform allows companies to comprehensively track metrics associated with large language models (LLM) and machine learning, generating alerts when these metrics fall below established levels. Additionally, the platform provides audit evidence, an essential component for compliance with current regulations.
As more companies adopt LLM, they face significant challenges, such as the need to monitor these models in real environments. Fiddler’s solution, called Guardrails, not only addresses this need but also offers response times under 100 milliseconds, ensuring safe deployment in customers’ work environments.
Fiddler’s commitment to responsible AI practices includes continuous monitoring for biases and equity. Their platform will alert in real-time about biased behaviors and automatically generate the necessary audit evidence to comply with regulatory standards.
A concrete example of Fiddler’s positive impact is reflected in its collaboration with Integral Ad Science (IAS), where Fiddler’s platform improved the monitoring of machine learning models and facilitated regulatory compliance, resulting in considerable cost savings and significant acceleration in product launch speed.
With a growing focus on data privacy, Fiddler has established a robust security framework that ensures compliance with data protection regulations. The company is committed to handling all interactions and data in accordance with current privacy laws and industry standards.
Looking to the future, Fiddler AI has plans to expand its leadership in AI observability, focusing on developing advanced capabilities and educational resources that will drive effective governance in using this technology, helping companies implement AI models that are not only cost-effective but also responsible and equitable for all.
via: MiMub in Spanish