1
Towards Artificial General Intelligence (AGI) in the
Internet of Things (IoT): Opportunities and
Challenges
Fei Dou1, Jin Ye2, Geng Yuan1, Qin Lu2, Wei Niu1, Haijian Sun2, Le Guan1, Guoyu Lu2, Gengchen Mai3,
Ninghao Liu1, Jin Lu1, Zhengliang Liu1, Zihao Wu1, Chenjiao Tan4, Shaochen Xu1, Xianqiao Wang5, Guoming
Li6, Lilong Chai6, Sheng Li7, Jin Sun1, Hongyue Sun5, Yunli Shao5, Changying Li4, Tianming Liu1, Wenzhan
Song2
Abstract—Artificial General Intelligence (AGI), possessing the I. INTRODUCTION
capacity to comprehend, learn, and execute tasks with human
cognitiveabilities,engenderssignificantanticipationandintrigue A. IoT Concept
acrossscientific,commercial,andsocietalarenas.Thisfascination
The Internet of Things (IoT) constitutes a transformative
extends particularly to the Internet of Things (IoT), a landscape
technological paradigm that envisions a world where physical characterized by the interconnection of countless devices, sen-
sors, and systems, collectively gathering and sharing data to objects and digital systems harmoniously interact to enhance
enableintelligentdecision-makingandautomation.Thisresearch the quality of human life and streamline various processes
embarks on an exploration of the opportunities and challenges [1],[2],[3].Inasimplifiedconceptualframework,IoTcanbe
towards achieving AGI in the context of the IoT. Specifically,
represented as a basic formula [4]:
it starts by outlining the fundamental principles of IoT and
the critical role of Artificial Intelligence (AI) in IoT systems.
Subsequently, it delves into AGI fundamentals, culminating in
IoT =Humans+Physical Objects(comprising devices,
the formulation of a conceptual framework for AGI’s seamless
integrationwithinIoT.TheapplicationspectrumforAGI-infused sensors,controllers,storage,etc.)+Internet
IoT is broad, encompassing domains ranging from smart grids, (1)
residential environments, manufacturing, and transportation to
environmental monitoring, agriculture, healthcare, and educa- This captures the essence of IoT’s interconnected nature,
tion.However,adaptingAGItoresource-constrainedIoTsettings where human interactions, the physical world’s components,
necessitates dedicated research efforts. Furthermore, the paper
and the vast expanse of the internet converge.
addresses constraints imposed by limited computing resources,
intricaciesassociatedwithlarge-scaleIoTcommunication,aswell
as the critical concerns pertaining to security and privacy.
Index Terms—Artificial intelligence (AI), Artificial General
Intelligence (AGI), Internet of Things (IoT), Machine Learning
(ML),LargeLanguageModels(LLM),FoundationModels(FM)
1FeiDou,GengYuan,WeiNiu,LeGuan,NinghaoLiu,JinLu,Zhengliang
Liu, Zihao Wu, Shaochen Xu, Jin Sun, and Tianming Liu are with the
School of Computing, University of Georgia, Athens, GA 30602, USA (e-
mail:{fei.dou,geng.yuan,wniu,leguan,ninghao.liu,jin.lu,zl8864,zihao.wu1,
shaochen.xu25,jinsun,tliu}@uga.edu).
2Jin Ye, Qin Lu, Haijian Sun, Guoyu Lu, and Wenzhan Song are
with the School of Electrical and Computer Engineering, University of
Georgia, Athens, GA 30602, USA (e-mail: {jin.ye, qin.lu, hsun, guoyu.lu,
wsong}@uga.edu)
3Gengchen Mai is with the Department of Geography, University of
Georgia,Athens,GA30602,USA(e-mail:gengchen.mai25@uga.edu)
4 Chenjiao Tan and Changying Li are with Agricultural and Biological
Fig.1. AnillustrationofIoTarchitectures.
EngineeringDepartment,UniversityofFlorida,Gainesville,FL32611,USA
(email:{c.tan,cli2}@ufl.edu)
5Xianqiao Wang, Hongyue Sun, and Yunli Shao are with the School Fig. 1 illustrates a fundamental depiction of the IoT system
of Environmental, Civil Agricultural and Mechanical Engineering, Univer- [5], [6]. Within this system, users and IoT devices are in-
sity of Georgia, Athens, GA 30602, USA (e-mail: {xqwang, hongyuesun,
terconnected via the Internet. Communication between users
yunli.shao}@uga.edu)
6Guoming Li and Lilong Chai are with the Department of Poultry and devices is facilitated by gateways and signals, which
Science, University of Georgia, Athens, GA 30602, USA (e-mail: {gmli, are transmitted and routed through signal receiving towers.
lchai}@uga.edu)
Additionally,thedatasharedamongtheseinterconnectedusers
7ShengLiiswiththetheSchoolofDataScience,UniversityofVirginia,
Charlottesville,VA22903,USA(e-mail:vga8uf@virginia.edu) and devices can be archived on the public cloud for later
3202
peS
41
]IA.sc[
1v83470.9032:viXra