Intelligent Informatics at Rutgers: Open Data & Artificial Intelligence

Slides

Poonam Soans – NJOIT Open Data Center

Rakesh Kumar – Benefits and Challenges of Generative AI

Policy Dynamics of Open Data and AI

Summary

by Julia Esguerra

The Open Data + AI forum was held just before Thanksgiving on November 17th, hosted by the Public Informatics at Rutgers University program in collaboration with The New Jersey Big Data Alliance, Heldrich Center for Workforce Development, Northeast Big Data Innovation Hub, and Rutgers University OARC. The well-attended virtual panel featured an insightful panel discussion where we heard from four speakers from a diverse range of backgrounds, covering topics ranging from open data initiatives and technological advances to societal reactions and the media’s role in the AI landscape. The forum started with welcoming remarks from Dean Stuart Shapiro, who expressed the boundless opportunities and considerable challenges in the intersection of open data and AI, highlighting the need for diverse disciplines to constantly recalibrate in response to technological impacts, and emphasizing the importance of building bridges between technical and non-technical perspectives to successfully balance the promises and risks of technological advances.

The first speaker was Poonam Soans, the state of New Jersey’s Chief Data Officer, who highlighted New Jersey’s open data portal, emphasizing its standards and best practices for data sharing and transparency. The portal, data.nj.gov, includes diverse datasets such as public employee payroll records, performance budgeting, state expenditures, and more. Poonam discussed the advantages of open data, emphasizing its one-time effort for data input, broad accessibility across departments and to the public, and the potential for citizens to build applications on top of the data. She also stressed the importance of making open data AI-ready by establishing a solid foundation of data governance.

Dr. Rakesh Kumar provided a fascinating presentation, featuring an interesting video on the potential of AI technology in robotics, where he demonstrated how a robot could search for a person in a room hiding behind furniture! He focused on the technological advances introduced by Large Language Models (LLMs), emphasizing their state-of-the-art performance in Natural Language Processing (NLP), diverse task capabilities, and potential for robots to navigate environments using 3D scene graphs. Dr. Kumar acknowledged challenges such as biased results and lack of explainability in AI. He emphasized the dynamic nature of knowledge, suggesting that there is hope and opportunity in what can be learned tomorrow.

Prof. Clinton Andrews then discussed societal reactions to AI advancements, particularly the challenges posed by chatbots like ChatGPT. He questioned whether society would do the right thing, highlighting the need for regulation and ethical considerations in the emerging AI industry. Prof. Andrews delved into the role of individuals, organizations, and government in shaping ethical practices in AI. He emphasized personal ethics, professional norms, and the influence of government policies in creating a responsible AI landscape. The discussion also touched on the role of public data in resolving issues of data ownership.

Bringing an integrated and multifaceted perspective, Rachel Rosenthal provided an enlightening perspective on the media. She explored the media’s place in the AI ecosystem, describing the traditional newsroom structure. She highlighted the challenge shared by the media and policymakers in understanding the scope of AI-related problems in order to propose effective solutions. Rachel discussed the difficulty of projecting ahead in the complex AI landscape. She emphasized the slow and messy process of policymaking, acknowledging the ongoing struggle to strike a balance between preserving innovation and protecting citizens in the context of AI.

The forum concluded with questions from the audience and an open discussion between the panelists moderated by Prof. Jim Samuel. In the panel discussion, speakers identified Large Language Models (LLMs) as a current AI technology with significant potential impact, emphasizing their role in providing easy access to knowledge and the ability to distill information. While cautioning about the need for expert user discernment, they acknowledged the potential for LLMs to complement human capabilities rather than serving as substitutes. Regarding the shaping of society, the panel highlighted both hopes and fears associated with AI and open data, emphasizing the potential for better communication and knowledge distillation and the need for adaptability. On the topic of AI regulation, the speakers expressed the view that regulation is necessary, envisioning a process of building upon existing frameworks and acknowledging the potential for a long and messy but innovative journey. They advocated for developing regulation from the bottom up, emphasizing AI’s life-changing use cases.

• This event was hosted by the Master of Public Informatics program at Rutgers University, in collaboration with our partners, NJBDA, Heldrich Center, NEBigData InnovationHub & Rutgers OARC
• This discussion was inspired by Bloustein research on the Garden State Open Data Index (GSODI).
• Please email informatics@ejb.rutgers.edu for additional information.