putilov_denis - stock.adobe.com
As legislators around the world focus on regulating AI systems, some companies are releasing tools to help enterprises govern their AI projects.
Monitaur, an AI governance software company, on Tuesday announced general availability of GovernML, the latest addition to its Machine Learning Assurance (MLA) platform.
Monitaur's MLA platform is part of a framework of products and tools that enables enterprises to keep the documentation, evidence and governance of AI projects in one place.
The release of GovernML comes as China became one of the first countries to require businesses to provide explainable and transparent AI algorithms on March 1st.
China is not the only country looking to regulate AI systems. Last year, the European Union released a draft of its Artificial Intelligence Act, which focuses on how specific organizations should utilize their AI systems. For its part, the U.S. is also considering ways to regulate AI, with the Federal Trade Commission publishing a blog post last year encouraging organizations to aim for truth and fairness in their use of AI.
While government entities are focused on regulating AI systems, many organizations are still learning how to govern their AI projects, according to Monitaur.
What is GovernML?
GovernML is built to help enterprises that have no idea how to govern their AI projects or implement risk management for their projects, Monitaur said.
"Before you start monitoring or testing or validating, every business should first establish: 'What are the things I am worried about? And what are the controls and processes that might mitigate those?" said Anthony Habayeb, CEO of Monitaur.
GovernML enables enterprises to build their risk policies for responsible AI. The tool helps enterprises with policy management, technical monitoring, testing and human oversight of their AI projects.
Tools like GovernML are the first generation of AI governance tools, said Kashyap Kompella, an analyst at RPA2AI Research.
Kashyap KompellaAnalyst, RPA2AI Research
"Even in the absence of regulation, many companies are committing themselves to responsible AI, and governance tools can help such companies right now," Kompella said.
Implementing AI governance tools
However, implementing and adopting these tools is not an easy process.
"Organizations are hesitant to introduce new tools to their enterprise tech stack because tool sprawl is a concern," Kompella said, referring to the practice of expanding your tool stack to the point of hindering rather than helping your organization's progress. "The challenge would be to integrate governance tools in the enterprise AI and data science workflows."
Despite this hesitancy, AI governance functionality will start to appear in MLOps products, general risk management platforms and software assurance tools in the next three to five years, he added.
"As the adoption of AI increases, we expect these tools to evolve as well to accommodate the needs of different organizations," Kompella said.
GovernML is a web-based, SaaS application. Pricing for the tool varies based on the number of models, workspaces and decision models.