Webinar at UMass Amherst Explores AI Policy Insights
On October 8, the University of Massachusetts Amherst’s College of Social and Behavioral Sciences conducted a noteworthy webinar featuring Alexander Macgillivray, the former U.S. Deputy Chief Technology Officer under the Biden administration.
His discourse revolved around his pivotal role in shaping artificial intelligence policy within the United States.
Macgillivray’s presentation, entitled “The Past and Future of AI and Regulation: A Practitioner’s View,” was the centerpiece of the UMass Social Science Series titled “AI and Us: Rethinking Research, Work, and Society.”
This series is designed to convene preeminent voices from academia and industry to engage in discussions, workshops, and hands-on events that scrutinize artificial intelligence’s implications in the social sciences.
He commenced his address by outlining the framework for an AI Bill of Rights—this initiative marked a significant regulatory stride taken by him and the Biden administration. Finalized in October 2022, prior to the emergence of ChatGPT, the document seeks to guarantee that AI systems are safe, effective, and accountable.
The Blueprint for the AI Bill of Rights aims to serve as a safeguard against potential threats posed by artificial intelligence to democracy, equity, and essential resources.
It is predicated on five guiding principles: the establishment of safe and effective systems, protections against algorithmic discrimination, data privacy, the necessity for notifications and explanations, and the consideration of human alternatives and fallback options, as outlined by the Office of Science and Technology Policy.
MacGillivray noted that there was an immediate surge of interest surrounding the implementation of these regulations. The bill notably resonated with public sentiment.
An article published by IEEE corroborated this notion, emphasizing that other AI researchers concurred that the linguistic choices within the bill clearly delineate AI governance as a critical civil rights matter warranting enhanced protections.
Further into his presentation, MacGillivray examined President Biden’s Executive Order 14110, which articulated the administration’s policy aspirations regarding artificial intelligence.
Although he was no longer serving in government, he provided insightful observations on the administration’s strategic approach.
“From an external vantage point, one could discern that the Biden administration grappled with the dual challenge of capitalizing on the benefits while mitigating the associated harms,” MacGillivray stated.
He elaborated that the executive order attempts to simultaneously navigate myriad questions pertaining to the risks and advantages of emerging technologies.
He also highlighted that while the Biden administration established restrictions governing AI chip exports and the locales for large-scale AI model training, such limitations had been relaxed during the Trump administration.
This reversal occurred by abrogating the Biden Administration’s Framework for Artificial Intelligence Diffusions, also known as the AI Diffusion Rule, which sought to regulate global transactions involving artificial intelligence and enforce licensing requirements for exports.
The Trump administration retracted this framework in May 2025, immediately prior to its scheduled implementation. The U.S. Department of Commerce asserted that this reversal was imperative to prevent the degradation of diplomatic relations with numerous countries.
Executive Orders 14179 and 14141, both instituted by the Trump administration, concentrate on augmented infrastructure investment, easing regulations on power plants and data centers, as well as limiting governmental reliance on ideologically biased AI systems.
MacGillivray remarked that although the Executive Order aimed at curtailing the use of ideologically driven AI received approval, to his knowledge, proper implementation remains elusive.
“It appears that the government continues utilizing Elon Musk’s AI, which is notably engineered to ensure ideological consistency with his views,” he remarked.
Grok, a generative AI chatbot developed by Musk, saw its accessibility to all federal agencies confirmed through the Grok for Government agreement approved by the Trump administration in September 2025, as reported by the General Services Administration.
This chatbot has faced scrutiny, as analyses by the New York Times have pointed out that Musk has programmed the AI to embody his political preferences.
Macgillivray cautioned against a prevailing tendency in the AI policy realm where individuals forecast a singular future for artificial intelligence and tailor their policy suggestions accordingly.
“When divergent scenarios are disregarded, discussions often become disjointed, rooted in fundamentally differing assumptions,” he lamented. He advocated for a holistic consideration of all conceivable futures for AI while formulating adaptive policy solutions.
He concluded with a call for a shift in focus toward recognizing the inherent harms of AI, which would mitigate reliance on hypotheticals. Emphasizing the necessity for policymakers to deepen their understanding of technology, he stated that involving knowledgeable individuals in the process is critical.
“It’s essential to recruit talented individuals into government roles. They must be adequately trained,” MacGillivray insisted. “The government should actively experiment with these technologies, develop them responsibly, and evaluate their efficacy to enhance regulatory efficiency.”
Source link: Dailycollegian.com.
