Artificial Intelligence Safety and Risks

Global Leaders Take Action

By Roman Z. Matatov, Senior Manager, Robert Gaines, Director, and Michael Corcione, Partner

President Biden recently issued an executive order on Safe, Secure, and Trustworthy Artificial Intelligence. This was the culmination of months of meetings among the White House, industry groups and artificial intelligence (AI) experts designed to “advance a cohesive and comprehensive approach to AI-related risks and opportunities.”

As artificial intelligence capabilities grow and its application becomes more democratized, safeguards for its proper and effective use become more apparent and necessary. This executive order includes the following primary elements.

New Standards for AI

New standards would require developers of AI systems to share their safety test results and other critical information with the U.S. government to be incorporated into a national risk model. This would help ensure systems are safe, secure and trustworthy before they’re released to the public and require the development of standards, tools, testing and verification programs:

  • NIST (National Institute of Standards and Technology) will be tasked to develop new risk models and standards for pre-release testing,
  • DHS (Department of Homeland Security) will be responsible for developing an AI Safety and Security Board, and
  • Both the DHS and DOE (Department of Energy) will be tasked with addressing AI threats to critical infrastructure.

AI-enabled fraud and deception will be countered by establishing standards and best practices for differentiating between AI-generated content and official content with the use of digital watermarks.

The Responsible Use of AI

To ensure AI is not a detriment to the U.S. public, the executive order promotes the responsible use of AI by protecting against using AI to engineer dangerous biological materials through new standards for biological synthesis screening. Additionally, cybersecurity programs will be adapted to develop AI tools to find and fix vulnerabilities in critical infrastructure, as well as develop more effective methods for ensuring the privacy protections of consumers. The United States military and intelligence community are instructed to use AI safely and ethically and are charged to counter any adversarial use of AI.

Protecting American’s Privacy

The executive order also calls for Congress to pass bipartisan data privacy legislation to protect Americans’ privacy from risks exposed by AI. The goal is that AI research, technologies and tools used in AI systems and their development should preserve data privacy.

Advancing Equity and Civil Rights

The executive order ensures AI is not abused in Federal programs by providing guidance about its usage in benefits programs, the criminal justice system and areas where algorithms can be used to exacerbate discrimination. The executive order also looks to help advance research in healthcare and education through AI-enabled tools.

The executive order reflects the government’s efforts to control how AI evolves so that its application is used to maximize its benefits while establishing safeguards to ensure its use does not become a detriment to users, businesses and the country.

Promoting Innovation and Competition

The executive order also ensures the U.S leads in AI innovation and competition. It aims to level the playing field for businesses in the AI ecosystem by providing small businesses with access to new assistance and resource programs and tools for researchers and students for AI research in healthcare and climate science. The order also wants to expand the ability for highly skilled workers to work and stay in the U.S.

Global Attention to AI

On November 1, 2023, the United States joined 28 other governments, attending the UK-hosted AI Safety Summit, in publishing the Bletchley Declaration. This declaration shares many of the elements of President Biden’s executive order, highlights the fact that AI deployment is now a worldwide issue and calls on the international community to cooperate in addressing the promise and potential peril of AI. This declaration draws attention to the aim of realizing “human-centric” AI benefits, including progressing the United Nations Sustainable Development Goals. I It also acknowledges the many “inherently international” risks that must be managed through evidence-based engagement, collaborating on de-risking policies and alignment among public and private sectors.

Future Developments

PKF O’Connor Davies is committed to remaining attentive to AI and will continue to share information about AI developments, including considerations for preparing for the broad impact AI is expected to have on individuals, organizations and industries.

Contact Us

Our team of specialists is here to help. For more information, please contact your PKF O’Connor Davies client engagement partner or:

Senior Manager
240.534.2797 |

Director, Cybersecurity and Privacy Advisory
425.518.1974 |

Michael Corcione
Partner, Cybersecurity and Privacy Advisory