Designing Proactive Generative AI Systems with Autonomous Agents: A Proposal for a Paradigm Shift from Reactive Prompt-Based Models
DOI:
https://doi.org/10.22399/ijcesen.3895Keywords:
Proactive AI, Autonomous Agents, Generative Intelligence, Human-AI Interaction, Context-Aware SystemsAbstract
The article introduces a conceptual framework for proactive generative AI systems augmented by autonomous agents, which anticipate and act on user needs without explicit prompts. It addresses limitations of reactive models by proposing an architecture that integrates AI agents for contextual awareness and decision-making. The framework demonstrates improvements in user interaction time for email tasks and precision in healthcare risk prediction compared to reactive systems. Technical aspects including agent coordination mechanisms, privacy considerations, and user trust are explored alongside applications in healthcare, productivity, and education. The proposal acknowledges limitations such as reliance on simulated data while providing guidance for future implementation rather than reporting on a deployed system.
References
[1] T. Brown et al., (2020). Language models are few-shot learners, in Proc. NeurIPS, 1877-1901.
[2] R. Rombach et al., (2022). High-resolution image synthesis with latent diffusion models, in Proc. CVPR, 10684-10695.
[3] L. Suchman, (1987). Plans and Situated Actions. Cambridge, U.K.: Cambridge Univ. Press.
[4] E. Hutchins, (1995). Cognition in the Wild. Cambridge, MA, USA: MIT Press.
[5] S. Russell and P. Norvig, (2020). Artificial Intelligence: A Modern Approach, 4th ed. Boston, MA, USA: Pearson.
[6] J. Wei et al., (2022). Chain-of-thought prompting elicits reasoning in large language models," in Proc. NeurIPS, 24824-24837.
[7] M. Chen et al., (2019). Gmail smart compose: Real-time assisted writing, in Proc. KDD. 2287-2295.
[8] A. Pejovic et al., (2023). Anticipatory computing: From mobile to IoT, IEEE Pervasive Comput., vol. 22(2), 25-33.
[9] S. Amershi et al., (2019). Guidelines for human-AI interaction, in Proc. CHI, 1-13.
[10] N. Vlassis, (2007). A concise introduction to multiagent systems and distributed AI, Synthesis Lectures AI, vol. 1(1), 1-71.
[11] S. LaValle, (2006). Planning Algorithms. Cambridge, U.K.: Cambridge Univ. Press.
[12] P. Christiano et al., (2017). Deep reinforcement learning from human preferences, in Proc. NeurIPS, 4299-4307.
[13] A. Vaswani et al., (2017). Attention is all you need, in Proc. NeurIPS, 5998-6008.
[14] S. Hochreiter and J. Schmidhuber, (1997). Long short-term memory, Neural Comput., vol. 9(8), 1735-1780.
[15] Y. Liu et al., (2020). RoBERTa: A robustly optimized BERT pretraining approach, in Proc. ICLR.
[16] C. Dwork, (2008). Differential privacy: A survey of results, in Proc. TAMC, 1-19.
[17] A. Datta et al., (2016). Algorithmic transparency via quantitative input influence, in Proc. IEEE Symp. Security Privacy, 598-617.
[18] W. Cohen, (2009). Enron email dataset, Carnegie Mellon Univ., Pittsburgh, PA, USA, Tech. Rep.
[19] A. Goldberger et al., (2000). PhysioBank, PhysioToolkit, and PhysioNet, Circulation, vol. 101(23). e215-e220.
[20] D. Kahneman, (2011).Thinking, Fast and Slow. New York, NY, USA: Farrar, Straus and Giroux.
[21] J. Wu et al., (2023). AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation, arXiv preprint arXiv:2309.03409.
[22] OpenAI, (2024). GPT Agents: Autonomous Task Execution with LLMs, OpenAI Blog.
[23] Microsoft, (2023). Jarvis: LLM-Powered Autonomous Agents, Microsoft Research.
[24] IEEE, (2023). Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems, IEEE Standards Association.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 International Journal of Computational and Experimental Science and Engineering
This work is licensed under a Creative Commons Attribution 4.0 International License.
