How should we deploy the mighty powers of artificial intelligence to fight poverty, improve health and wellness, and expand opportunity? Carefully. Very carefully.
For nearly 20 years, Benefits Data Trust (BDT) has tapped the power of data and technology to connect people facing tough times to life-saving and life-improving public assistance and to help public servants and others make America’s public benefits programs more effective, equitable, and dignified. We are committed to ensuring that America’s social safety net serves everyone better — that means allowing anyone to apply for benefits in under 20 minutes, find out if they will receive them within 24 hours, and continue receiving help as long as they are eligible and want it. Ultimately our goal is to eliminate America’s $80 billion benefits access gap in the next 10 years, so millions of eligible people in need can easily access available assistance.
Achieving this requires making these programs more intelligent, including harnessing the potential of AI. To help achieve that, BDT is announcing the launch of a Trustworthy AI and Human Services Learning Hub, made possible through the generous support of The Rockefeller Foundation, Patrick J McGovern Foundation, and others. This Learning Hub will provide a sandbox environment for collaboration between government, the private sector, academia, labor, civil society, front-line agency workers and benefits recipients. This effort seeks to:
- Define and build consensus toward trustworthy AI and human services drivers, priorities, and principles;
- Test the socio-technical effects of applying AI across a spectrum of rights-affecting and public-impacting use cases;
- Inform policymaking, public administration, product innovation, and trustworthy AI discourse through continuous dialogue among diverse stakeholders; and
- Build a roadmap to trustworthy AI for human services and the vehicles to get there, as part of delivering efficient, dignified, and equitable access to public benefits.
As BDT’s chief digital officer Stephen Rockwell has said: “Well-governed data based on real world interactions is vital to creating reliable and accurate models in the development of generative AI and other forms of AI applications. These applications and modeling are the keys to closing the gap between people who are eligible for benefits and those enrolled in benefit programs, and fundamentally transforming our nation’s $1 trillion plus social safety net system with profound implications for households, small businesses, nonprofits, and government.”
"BDT is launching a Trustworthy AI and Human Services Learning Hub that will provide a sandbox environment for collaboration between government, the private sector, academia, labor, civil society, front-line agency workers and benefits recipients."
BDT’s use of AI includes powering a digital advisor helping students apply for college financial aid and streamlining critical but repetitive back office functions so our colleagues have more time to serve the people who call our Benefits Centers to get help navigating the application process. We are also testing and applying AI’s capabilities to expand proactive and culturally relevant benefits access; recommend social support interventions that lead to successful outcomes; put BDT’s benefits access expertise in the hands of frontline workers and allies who people know and trust locally across the country; and bring well-tailored, human controlled automation to the benefits application and renewal process.
But getting the most out of AI requires facing head on some formidable ethical, technical, operational, and policy challenges to ensure that the harms AI can cause do not negatively impact the lives of people most in need and the programs designed to help them. Getting this right will be an all-hands, iterative effort. The Trustworthy AI and Human Services Learning Hub will help jumpstart this collaborative work.
We encourage you to contact us if you are interested in participating in this work.
Participate in the Learning Hub
Are you interested in how AI should be used by government human services? Please sign up to express interest in participating in the Trustworthy AI and Human Services Learning Hub.
Watch this video to hear more about BDT’s vision for an intelligent human services and benefits access system.