TY - GEN
T1 - AI-Governance and AI for Governance
T2 - 29th Annual Americas Conference on Information Systems: Diving into Uncharted Waters, AMCIS 2023
AU - Eisenberg, David
AU - Abhari, Kaveh
N1 - Publisher Copyright:
© 2023 29th Annual Americas Conference on Information Systems, AMCIS 2023. All rights reserved.
PY - 2023
Y1 - 2023
N2 - Generative Artificial Intelligence (AI) governance of political economy could allow for efficient allocation of resources by interpreting vast amounts of data and advanced analytical capabilities to predict market needs and forces. While this might resemble a central planning board, it differs in that an AI system is adaptive, learning from market fluctuations and making adjustments accordingly, while also reducing the impact of human corruption and self-interest. The opportunities and potential drawbacks of such a system are vast, hold significant economic and societal consequences, and warrant thorough discussion of its key topics: 1. Unanticipated decisions of evolving AI: As AI systems grow and learn, there is a possibility that they could make unexpected decisions. Some ways to mitigate these risks include: a) Continuous monitoring and evaluation: Regularly assessing the AI's performance and decision-making can help identify potential issues and provide opportunities for adjustment. b) Human intervention: Ensuring that humans maintain the ability to intervene and override the AI's decisions can provide a safeguard against unforeseen consequences. c) Ethical guidelines: Establishing clear ethical guidelines for the AI system can help maintain its alignment with human values and societal goals. 2. There is a risk that an AI-governed economy could lead to a surveillance state. Ensuring transparency and accountability is crucial for avoiding this outcome. Potential measures to address this concern include: a) Open Source: Making AI's algorithms open-source could allow for public scrutiny and prevent the AI itself from unchecked authority. However, how do we balance the need for open access to code with the threat of human infiltration or hacking into a system intended to be free of undue influence? b) Public input: Allowing citizens to participate in decisions regarding data privacy and surveillance policies could help ensure that the AI's data gathering remains transparent and accountable. c) Oversight committees: Creating committees of diverse stakeholders, including experts, citizens, and representatives from industry, can help oversee the AI's decision-making and maintain a balance of power. 3. Ensuring a beneficial system for all: The four pillars for a successful AI-governed system could include: a) Fairness: Ensuring that the AI system distributes resources equitably among all citizens and does not favor any particular group or interest. b) Accountability: Implementing measures to hold the AI system accountable for its decisions, such as oversight committees and public input mechanisms. c) Sustainability: Designing an AI system to prioritize long-term societal and environmental goals. d) Transparency: Making the AI's decisions and algorithms transparent, allowing public scrutiny and input In conclusion, while AI governance of political economy and economic management poses risks, careful design, implementation, and ongoing oversight can help ensure that such a system remains fair, accountable, sustainable, and transparent. By addressing concerns around surveillance, unanticipated decisions, and evolving AI, we can create a FAST-designed system that benefits society as a whole.
AB - Generative Artificial Intelligence (AI) governance of political economy could allow for efficient allocation of resources by interpreting vast amounts of data and advanced analytical capabilities to predict market needs and forces. While this might resemble a central planning board, it differs in that an AI system is adaptive, learning from market fluctuations and making adjustments accordingly, while also reducing the impact of human corruption and self-interest. The opportunities and potential drawbacks of such a system are vast, hold significant economic and societal consequences, and warrant thorough discussion of its key topics: 1. Unanticipated decisions of evolving AI: As AI systems grow and learn, there is a possibility that they could make unexpected decisions. Some ways to mitigate these risks include: a) Continuous monitoring and evaluation: Regularly assessing the AI's performance and decision-making can help identify potential issues and provide opportunities for adjustment. b) Human intervention: Ensuring that humans maintain the ability to intervene and override the AI's decisions can provide a safeguard against unforeseen consequences. c) Ethical guidelines: Establishing clear ethical guidelines for the AI system can help maintain its alignment with human values and societal goals. 2. There is a risk that an AI-governed economy could lead to a surveillance state. Ensuring transparency and accountability is crucial for avoiding this outcome. Potential measures to address this concern include: a) Open Source: Making AI's algorithms open-source could allow for public scrutiny and prevent the AI itself from unchecked authority. However, how do we balance the need for open access to code with the threat of human infiltration or hacking into a system intended to be free of undue influence? b) Public input: Allowing citizens to participate in decisions regarding data privacy and surveillance policies could help ensure that the AI's data gathering remains transparent and accountable. c) Oversight committees: Creating committees of diverse stakeholders, including experts, citizens, and representatives from industry, can help oversee the AI's decision-making and maintain a balance of power. 3. Ensuring a beneficial system for all: The four pillars for a successful AI-governed system could include: a) Fairness: Ensuring that the AI system distributes resources equitably among all citizens and does not favor any particular group or interest. b) Accountability: Implementing measures to hold the AI system accountable for its decisions, such as oversight committees and public input mechanisms. c) Sustainability: Designing an AI system to prioritize long-term societal and environmental goals. d) Transparency: Making the AI's decisions and algorithms transparent, allowing public scrutiny and input In conclusion, while AI governance of political economy and economic management poses risks, careful design, implementation, and ongoing oversight can help ensure that such a system remains fair, accountable, sustainable, and transparent. By addressing concerns around surveillance, unanticipated decisions, and evolving AI, we can create a FAST-designed system that benefits society as a whole.
UR - https://www.scopus.com/pages/publications/85192921341
M3 - Conference contribution
AN - SCOPUS:85192921341
T3 - 29th Annual Americas Conference on Information Systems, AMCIS 2023
BT - 29th Annual Americas Conference on Information Systems, AMCIS 2023
PB - Association for Information Systems
Y2 - 10 August 2023 through 12 August 2023
ER -