Intel has published a white paper outlining the company's recommendations for the the U.S. National Strategy on Artificial Intelligence.
On February 11, 2019 the White House issued an Executive Order on Maintaining American Leadership in Artificial Intelligence.
"A comprehensive national AI strategy would earmark funding and resources for AI research and development, outline clear goals and accountability mechanisms, identify and remove barriers, drive public and private development and adoption of AI, and outline a program to mitigate negative or unintended consequences," Intel said.
Intel's recommendation for a national AI strategy lays out four key responsibilities for government. Within each of these areas, the company is proposing actionable steps. Some highlights are provided below, and Intel's full white paper is available here.
Intel calls for sustainable and funded government AI research and development in order help advancing the capabilities of AI in areas such as healthcare, cybersecurity, national security and education, but there need to be clear ethical guidelines. The company calls for
specific funding commitments to AI research and development, starting with a study to determine the areas with the most potential for deployment of AI — including socially beneficial initiatives such as climate change and education.
The development of responsible government policies for AI can help the government’s efforts to engage the private sector and inform the development of international norms, according to Intel.
International cooperation and data interoperability standards should be also supported, in order to ensure intellectual property protections, Intel says.
Intel also calls the U.S gonvernment to invest in the development of a diverse workforce that creates AI and uses AI. Continuous education is the most widely agreed-upon approach to developing a workforce capable of creating AI systems. A “National Service” study would examine how a broad-based public/private partnership network of national service opportunities might alleviate potential job loss, while rebuilding infrastructure and developing skills in the use of technology, Intel says.
As the more data that is available, the more “intelligent” an AI system can become. However, Intel says that increased access to reliable data must be coupled with privacy protections and comprehensive ethical measures. "The adoption of comprehensive federal privacy legislation, and policies that require accountability for ethical design and implementation, are critical to ensuring protections are in place for efficient data sharing practices that mitigate potential individual and societal harm," according to Intel.
The company also calls for increased access to government data sources in machine readable formats across borders and jurisdictions will provide AI practitioners with a wide variety of data to help train AI systems. The adoption of international data interoperability standards will help speed the development and adoption of AI systems, and increase the efficiency of data sharing processes.
The creation of a legal and policy environment that supports AI will also lead to responsible development and use of AI. Intel proposes,
where necessary, the regulation on principles rather than regulating specific algorithms. It says that expanding general legal principles to AI is needed, along with protection of innovation and IP globally. This would help companies avoid requiring to transfer technology or IP as a condition of doing business.