Biden’s executive order on AI offers overdue ‘guardrails’ to protect privacy and reduce bias, this gov contractor says

Artificial intelligence is not new technology. Computer scientists were teaching programs to play checkers as far back as the 1950. But practical use cases through tools like ChatGPT, GitHub Copilot and Otter.ai have made it accessible to the general public in recent years.

Last month, President Biden issued an executive order on AI, establishing new standards for AI safety and security that protects Americans’ privacy, advances equity and civil rights, and promotes innovation and competition.

The implementation of guardrails around the use of AI is long overdue. These guardrails are necessary to reduce exploitation of people.

Tools like ChatGPT are driving a demand for AI integration, but it is critical we have policies and oversight to ensure that data collected and used for AI does not violate individual privacy and safety.

Like many private sector industries, the government sector has shown an increased interest in using AI more to automate work, improve the speed of processes or support predictive analytics. AI is already being used across many government agencies. At Fearless, we have integrated AI in our work with agencies such as Centers for Medicare & Medicaid Services and Department of Veterans Affairs.

It’s exciting to think about all the use cases where AI could improve efficiencies and access to government services. With AI tools, the FDA can anticipate the impact of drugs with predictive analytics, while the process for veterans to access benefits can be expedited, just to name a few.

However, AI has become a buzzword and organizations need to be prepared to ask the right questions to determine if the services offered are effective and appropriate. Simply including AI in a capabilities statement does not clearly establish exactly what ways a company can use AI to support an organization.

It is critical organizations understand what to look for and the questions to ask when evaluating vendors to support the development and integration of AI.Ravi Gourineni Fearless

The executive order outlines several ways the government plans to develop and implement protections around AI. The Departments of Energy and Homeland Security will address AI systems’ threats to critical infrastructure with the support of the National Institute of Standards and Technology, setting rigorous standards for extensive red-team testing to ensure safety before public release. There is a commitment to advancing a cybersecurity program to find and fix vulnerabilities, and developing guidance for content authentication and watermarking to clearly label AI-generated content.

These are all important steps. It is not clear though how quickly all of this will roll out or how it will be monitored and regulated. Therefore, it is critical organizations understand what to look for and the questions to ask when evaluating vendors to support the development and integration of AI.

In addition to policies and procedures, we need the right workforce to support AI and bring unbiased AI innovation to government services. This will require intentional efforts to build a more diverse tech talent pipeline, especially within federal contractors, and catalyze commercial growth of companies led by people of color.

Baltimore was recently designated one of the 31 inaugural federal tech hubs and AI is at the forefront. Fearless will play a critical role in this as a partner with the Greater Baltimore Committee. We will leverage Baltimore’s current strengths that focus on predictive technologies to improve health and wellbeing at the individual, community, and national level, while supporting the commercialization and scaling of this core technology to create jobs and wealth and mitigate disparities in the region.

Through this, we will incubate 25 AI technologists and companies, and we will train 50 workers from underrepresented and nontraditional backgrounds to work in AI and data science. Both of these strategies will result in the increase of new minority-owned companies and the skills training of underrepresented communities that are integral to ensuring there is broader representation in the future of AI.

This new executive order ensures responsible and effective government use of AI, expanding services and cutting costs. It requires the hiring of data professionals which will improve education across agencies and a more comprehensive understanding of AI and its capabilities, and accelerate the speed by which agencies can acquire AI products and services.

The executive order is a vital next step to regulating and monitoring the use of AI to protect and secure our nation, but much work is left to be done to implement, and importantly, enforce the actions outlined.

We commend the administration’s commitment to fostering responsible AI while ensuring its safety and security, and we are enthusiastic about contributing to this mission.