New report provides detailed steps to mitigate the risk of AI bias

WASHINGTON – June 8, 2021 – Artificial intelligence (AI) is helping people achieve incredible outcomes. Whether it’s improving healthcare or helping achieve sustainability goals, emerging AI technologies are unlocking innovation and economic growth. As AI is used in ways that will have increasingly consequential impacts on people’s lives, there is an urgent need for policymakers and industry leaders to align around best practices for mitigating the potential risks of AI bias.

Today, BSA is calling on governments to pass legislation to require private sector companies to perform impact assessments on high-risk uses of AI technologies. To aid governments in this effort, BSA today released Confronting Bias: BSA’s Framework to Build Trust in AI.

The framework details how organizations can perform impact assessments to identify and then mitigate risks of bias that may emerge throughout an AI system’s lifecycle. With more than 50 diagnostic statements specifying actions for companies to take, covering the Design, Development, and Deployment of an AI service, this is a first-of-its-kind framework.

This new framework is just one component of effective AI risk management. BSA also calls on governments around the globe to implement regulations and legislation to require impact assessments in order to minimize the potential of bias in high-risk uses of AI. BSA’s Framework to Build Trust in AI can serve as a resource for policymakers and industry leaders alike as they grapple with how best to leverage the benefits of AI without perpetuating existing inequalities.

“Now is the time for industry to step forward and work with policymakers to pass legislation to address risks of AI bias, and BSA will help lead this effort. Companies and governments alike should use BSA’s AI Risk Management Framework as a playbook for building trust and transparency at every point in the AI lifecycle, from design to deployment,” said Victoria Espinel, President and CEO at BSA | The Software Alliance. “BSA’s members are committed to finding and implementing best practices to keep AI systems from producing unintentional negative consequences. Policymakers in the European Union have already begun work on legislation to regulate AI, and we will continue to work with the EU, with leaders in the US, and with policymakers around the globe to build the right approach and pass it into law.”

“AI has the potential to reshape industries and improve quality of life around the globe. But, in the absence of key safeguards, AI can also create feedback loops that may entrench and exacerbate historical inequities,” said Christian Troncoso, Senior Director, Policy at BSA. “Like cybersecurity and privacy, managing the risks of AI bias requires an organizational commitment that extends throughout a system’s lifecycle. Organizations can use the BSA Framework to enhance trust in their AI systems through risk management processes that promote fairness, transparency, and accountability.”

BSA’s Framework to Build Trust in AI sets forth a risk management framework that organizations can use to perform impact assessments to identify and mitigate potential risk of bias that may emerge throughout an AI system’s lifecycle. Additionally, the Framework:

• Sets out the key corporate governance structures, processes, and safeguards that are needed to implement and support an effective AI risk management program; and
• Identifies existing best practices, technical tools, and resources that stakeholders can use to mitigate specific AI bias risks that can emerge throughout an AI system’s lifecycle.

Explore the full BSA Framework to Build Trust in AI here.