Exploration on the Fusion of Artificial

Intelligence and Blockchain

Chongxiao Qu, Lei Jin, Yongjin Zhang, Changjun Fan(B) , Jinqi Chu, and Shuo Liu

The 52nd Research Institute of China Electronics Technology Group Corporation,

Hangzhou 310012, China
cethiker@[Link]

Abstract. Blockchain and artificial intelligence (AI) are the two leading tech-
nologies in the world, which promote the development of the Fourth Industrial
Revolution. Although there is lots of research on AI and blockchain separately,
few studies focus on the fusion or integration of these two. To address this gap,
this study aims to characterize the necessity and benefits of the fusion of AI and
blockchain. First, we introduce AI and blockchain technologies roughly. Then,
we analyze the general fusion mechanism for AI and blockchain and its necessity.
Next, we carry out the critical point analysis one by one for the fusion of AI and
blockchain. The detailed analysis shows that AI and blockchain technologies can
enhance each other and it is wise to integrate their benefits.

Keywords: AI · Blockchain · Smart contracts · Data · Computing power ·

Algorithms

1 Introduction
Nowadays, we have witnessed that AI and blockchain are the two leading technologies
worldwide, and they have attracted wide attention in both academia and industry [1].
Among them, AI technology is an essential driving force to promote the development
of the real economy, which has brought numerous open innovations and made itself
rise to the national strategic level. At the same time, emerging blockchain technology
has been listed in the national informatization planning of countries around the world,
whose technological development promotes a new round of industrial innovation and
constantly improves the industrial advantages in emerging fields.
As two technological trends, blockchain and AI have their own distinctive fea-
tures, and the focus of these two, as well as their development, also varies. In essence,
blockchain technology generates and stores data through consensus algorithms, operates
data through smart contracts, and ensures data security through cryptography. Blockchain
technology has the characteristics of decentralization, ensures data not be tampered with,
and focuses on maintaining accurate records, authentication and execution. While AI
technology involves three key points, namely data, algorithms and computing power. It
facilitates mining, evaluation and understanding of certain patterns in real-life industrial
datasets, makes decisions based on them, and results in autonomous interactions. On one

© The Author(s) 2023

S. Kadry et al. (Eds.): BDEIM 2022, AEBMR 233, pp. 156–163, 2023.
[Link]
 Exploration on the Fusion of Artificial Intelligence 157

hand, both blockchain and AI have common characteristics and technical areas, which
will be the key to their integrated development. On the other hand, each has its own
strengths and weaknesses, and there is a chance for them to complement each other. In
this paper, we try to explore the possibility of leveraging the benefits of blockchain and
AI together at different levels, e.g., data, algorithms, computing power, and so on. After
extensive analysis, we want to prove that, through their fusion, a disruptive impact will
be exerted on the technology industry in all areas of society.

2 AI and Blockchain Technology Overview

2.1 AI Technology Overview
The term “artificial intelligence” was proposed at the academic conference held by
Dartmouth University in 1956, when was deemed the starting point of global AI research.
According to Dr. Norvig, AI can be seen as an intelligent agent designed by human
beings, which refers to any system that perceives its environment and takes actions that
maximize its chance of achieving its goals [2]. AI is not a new term, but a research field
that has experienced decades of ups and downs. In March 2016, the world Go champion,
Li Shishi, was defeated by AlphaGo with a score of 1:4, which event has ignited the
world since then. In the following years, DeepMind Inc. Released several new versions
of AlphaGo robots, including AlphaGo Zero, AlphaGo Master, etc. which need less
training time and own higher intelligence. The resurgence of AI technology is due to the
support of intelligent hardware development and new breakthroughs in machine learning
(ML) and deep learning (DL).
ML is a technology that allows computers to reason without explicit programming,
which is the best way to develop high-level AI. Traditional ML technology relies on the
experience of domain experts to extract features. In contrast, DL technology that has
emerged in recent years is a representation learning method with multi-layer neurons,
which allows the machine to input raw data and automatically learn the required repre-
sentation from the data. It accomplishes this task by combining simple but non-linear
modules, in which each module transforms a lower-level representation into a higher and
more abstract representation. Due to this characteristic, DL is very good at discovering
complex structures in high-dimensional data.
Apart from the positive aspects, AI’s rapid progression also causes a series of con-
cerns. For example, AI has brought lots of trepidations, such as fake news with realistic
but fake photographs or voices, invasions of privacy etc. Data, computing power and
algorithms are AI’s three core elements. As the “raw materials”, multi-source, real-time,
massive and multi-type data are the basic support to ensure high-quality models trained
by intelligent algorithms. Computing power is the “booster” to achieve large-scale com-
puting of intelligent algorithms. The acceleration of computing power such as GPU is an
important infrastructure to promote the training and reasoning efficiency of intelligent
algorithms. And, algorithms are the “power” to realize intelligence, which solves prob-
lems through iterative computing and brings innovations to current technologies and
thinking mode. Therefore, data, computing power and algorithms are called the Troika
of AI. However, the cost of computing resources along with the high barriers to acquiring
data and talent could lead to the monopolization of AI power among the big players.
158 C. Qu et al.

2.2 Blockchain Technology Overview

Since the publication of the Bitcoin White Paper in 2008, blockchain as the underlying
technology of bitcoin has received extensive attention and discussion. According to
Wikipedia, a blockchain is essentially a growing list of records linked and protected using
cryptography [3]. The recorded unit is called a block, and each block contains a hash
pointer (as a link to the previous block), a timestamp, and transaction data. Blockchain
is inherently resistant to data tampering. It can be used as an open distributed account
book and can effectively and permanently record transactions between both parties.
All transaction records are traceable and verifiable. Simply put, blockchain is a set of
protocols and a set of specifications, rather than specific codes or projects.
Blockchain can be described from a narrow or broad perspective. From a narrow per-
spective, a blockchain is a chain data structure, which is composed of several data blocks
connected in sequence according to time. Its chain structure and the adopted cryptology
technology ensure that the data can’t be tampered with and forged. So, the blockchain
can be regarded as a distributed ledger. From a broad perspective, blockchain technology
is a pioneering distributed infrastructure and computing paradigm involving a variety
of technologies. The data structure, distributed consensus algorithms and cryptology
algorithms adopted by it provide guarantees for storage and verification, generation and
update, security and non-tampering of data.
In a broader sense, blockchain functions through an underlying framework which can
be subdivided into six independent but inseparable architecture layers, namely, the data
layer, the network layer, the consensus layer, the incentive layer, the contract layer, and
the application layer. The data layer manages the data record types and data structures,
and specifies the data composition of the block; The network layer provides networking,
message propagation, message verification and other functions for nodes in the net-
work based on decentralized technologies; The consensus layer enables peer nodes in
the decentralized network to reach a consensus on the consistency of the content; The
incentive layer provides rewards for the consensus nodes to ensure the normal operation
of the blockchain system; The contract layer encapsulates programmable application
codes such as automation script and smart contract; The application layer provides var-
ious applications, such as digital cryptocurrency represented by bitcoin. In some layers,
high-level intelligent computing techniques are needed.

3 Analysis on Fusion Mechanism for AI and Blockchain

Based on data, AI mines knowledge from data, while blockchain is essentially a data
storage method, or “hyper ledger”, which supports the realization of data intelligence.
And blockchain technology can ensure reliable data to a certain extent, share data under
the condition of protecting data privacy, provide high-quality data for AI modeling,
and thus improve the accuracy of the model. Therefore, the two technologies that are
closely related to data can be effectively combined to complement each other and achieve
technological upgrading (Fig. 1).
Blockchain technology has the characteristics of being distributed, open, transparent,
traceable, and difficult to tamper with. It can improve the authenticity, relevance and
effectiveness of data used in AI applications by combining multiple technologies, such
 Exploration on the Fusion of Artificial Intelligence 159

Fig. 1. Integration of AI and Blockchain [4].

as distributed data storage, point-to-point transmission, consensus mechanisms, and

encryption algorithms. Also, it can ensure the intelligence level of AI models in three
different aspects, i.e., data, computing power and algorithms, innovate the collaborative
mode and computing paradigm of AI technology, and build a new AI ecosystem. Based
on the blockchain, AI can realize decentralized intelligent joint modeling and provide
users with flexible computing capabilities to meet specific application needs.
The smart contract in a blockchain is essentially a piece of code that implements
a certain algorithm, which can be made more intelligent by AI technology. AI has the
characteristics of being intelligent and automatic. Through optimization and simulation,
AI algorithms can promote the natural evolution and data sorting of the blockchain,
effectively prevent the occurrence of bifurcation of linked nodes, and more efficiently
handle the operation of the blockchain. For example, AI algorithms can transform tra-
ditional contracts into smart contracts, effectively solving the problems of security, ease
of use and reliability. At the same time, decentralized learning systems such as “joint
learning” introduced by AI can be used to solve the problem of redundant information
on the blockchain, which can improve the extensibility and efficiency of the blockchain.
Given their interdependence with each other, There’s no reason not to believe it will
be a good idea to fuse AI and blockchain technologies. In the following section, we will
conduct a detailed analysis of their integration from different levels and aspects.

4 Key Points Analysis for Fusion of AI and Blockchain

4.1 Data Level Analysis

At present, massive data in the AI industry lacks a unified and efficient sharing mecha-
nism and management method. The poor maintenance of open-source datasets leads to
uneven data quality with the data being not centralized and unified. Additionally, a large
amount of data required for model training are mainly available only in the government
and large companies. The regulatory restrictions and high commercial threshold lead to
160 C. Qu et al.

poor data circulation and difficult access, which seriously restricts the pace of small and
medium-sized enterprises in developing AI applications. And, when an intelligent agent
is going to make a decision, it needs to obtain as much real-time data as possible to be a
reference. If there is not enough real-time data or the data is not real-time enough, then
this agent can only achieve limited intelligence. Moreover, due to the poor quality and
credibility of raw data and the invasion of personal privacy, the authenticity, timeliness
and credibility of the data in the big data trading market are greatly reduced.
The distributed database in a blockchain enables each node to share data efficiently
so that each participant on the network can access data, which provides more exten-
sive data access and a more effective data monetization mechanism for AI [5]. Firstly,
a decentralized data-sharing platform can be built based on blockchain technology. Its
updating and information recording are completed by the interaction of distributed enti-
ties, not by an authority. In the network, it is not allowed to tamper with data, modify
data or create false data at will. Secondly, data verification in a blockchain will promote
the establishment of cleaner and more organized personal data, thus providing smoother
data integration and forming a new data market. Based on homomorphic encryption,
zero-knowledge proof, differential privacy, and other technologies, data privacy secu-
rity protection in multi-party data sharing is realized. Finally, the incentive mechanism
and consensus mechanism based on blockchain technology have greatly expanded the
source channels of data acquisition. More open, shared and real-time data for analysis
will make the prediction and evaluation by AI more accurate, generate more reliable
algorithm models, and then improve the technical level of the whole AI model.

4.2 Computing Power Level Analysis

The cost of computing power is a big pain in the current AI industry. As data grows
exponentially and algorithms become more and more sophisticated, AI models have
reached the scale of hundreds of billions of parameters trained on trillions of training
data samples, which undoubtedly requires stronger computing power. It needs more than
one million yuan to purchase GPU and other hardware resources to build a computing
center, which is too expensive for most small and medium-sized enterprises, resulting in
the current situation of “insufficient computing power, high cost and difficult to obtain”.
According to an analysis report [6] released by OpenAI, since 2012, the amount of
computing required to create the most advanced AI system has increased 10 times every
year, and computing power has become a major bottleneck in the development of AI
applications.
Distributed computing in the blockchain can use idle GPUs of small, medium-sized
enterprises or individuals as computing nodes to share computing resources and pro-
vide computing power for AI modeling. For example, through blockchain technology
in Xunlei Inc., users can share idle resources such as bandwidth, storage and comput-
ing capacity through Xunlei cloud devices [7]. Although the computing power of each
device is very small, when the number of devices reaches a certain magnitude, the accu-
mulated computing power is also huge. Based on the distributed network, the blockchain
helps to build a decentralized AI computing infrastructure. It can realize the operation of
neural network models on the decentralized massive nodes distributed around the world,
 Exploration on the Fusion of Artificial Intelligence 161

make use of the idle computing resources of each node for computing, realize decen-
tralized intelligent computing, and change the traditional idea of constantly improving
the performance of equipment to improve computing power. In addition, through the
smart contract, computing nodes in the network can be dynamically adjusted according
to the computing amount of the user’s applications, so as to provide flexible computing
capacity to meet the user’s specific needs.
Mining the blockchain requires lots of computing resources and power. At present,
the annual power consumption of bitcoin is about 2.55 billion watts, almost equivalent to
the annual power consumption of some small countries. Applying AI techniques to POW
consensus mechanism and hash operation can greatly improve calculation efficiency,
thus saving power and energy. For example, Matrix, a start-up enterprise, uses AI to
combine POW and POS and adopts a hierarchical consensus mechanism. First, it uses
random clustering algorithms to generate multiple small clusters in the network and elects
representative nodes based on POS mechanism. Then, the elected nodes compete for
POW accounting rights. Compared with the competitive accounting method involving
all the nodes, it can greatly reduce energy waste. Besides, the intelligent system can
calculate the probability of a specific node performing tasks first, so as to remind the
miners to find other paths and reduce the total operation cost. Additionally, optimizing
energy consumption techniques in AI can also be applied to the blockchain, thereby
reducing the investment in mining hardware.

4.3 Algorithm Level Analysis

Nowadays, there is a large demand for AI technology in the market. However, at the
algorithm level, the threshold of AI algorithm research and development as well as the
requirement for talent teams are high. The current algorithms can only meet the needs
of a few enterprises. The development of personalized products has high technical and
financial barriers, which makes it extremely difficult for most enterprises to develop AI
independently. In such an information-explosive era, the construction and sharing of
information resources have become the need and necessity of the times. Now, advanced
AI development tools are only in the hands of a few people. For example, the program-
ming community created by independent researchers can’t be installed, configured and
run by the outside world, and their algorithms and models can’t be accessed. Information
island and interception have become the biggest barrier to resource sharing. Addition-
ally, most enterprises obtain open-source algorithms from the open-source community,
but there are many rules or version adaptation problems, which seriously limit the flexi-
bility and innovation of application development. When hosting or contributing code in
the open-source community, improper management of proprietary intellectual property
code may cause the risk of property disclosure.
With the characteristics of distributed collaboration of blockchain, a platform for
releasing ML tasks can be built. AI algorithms can be optimized, updated and main-
tained by multiple AI experts to realize algorithm sharing. Blockchain can bring an
open AI development mode to allow more people to participate and share AI resources,
which is conducive to improving resource utilization and AI’s long-term development
[8]. Additionally, an algorithm trading market can be built based on blockchain tech-
nology, allowing users to publish tasks and purchase algorithm models on the platform,
162 C. Qu et al.

encouraging developers, and greatly reducing the threshold of AI use under the condition
of ensuring intellectual property rights and privacy. The AI + blockchain can encourage
developers of AI models to share their development achievements.

4.4 Intelligence + Contract

The emergence of blockchain technology redefines traditional contracts, which further
details and deepens the connotation of smart contracts. Current smart contracts are
generally coded programs that have status, are event-driven, comply with certain protocol
standards, and run on the blockchain. It can process and operate the blockchain data in the
form of events or transactions under certain triggering conditions according to the code
rules, and thereby control and manage the digital assets of the blockchain network. As an
embedded programmed contract, it can be built into any blockchain data, transactions,
and tangible or intangible assets to form a programmable controlled software-defined
system, market and assets.
A smart contract is only a system composed of transaction processing modules and
state machines. It exists only to enable a group of complex digital commitments with
trigger conditions to be correctly executed according to the will of participants [9]. In
addition, smart contracts are not legally binding, nor functionally smart. In the case that
commercial activities really need to actually sign contracts, smart contracts can’t realize
the functions given by their names in practice and theory. The smart contract code itself
also lacks the basic elements of real contracts, such as terms, conditions and dispute
resolution. Smart contracts still have a long way to go before they are truly legally
binding. Smart contracts have security problems easily exploited by hackers, and human
intervention makes their vulnerability unable to be repaired online, ending in more
serious problems or chain reactions [10]. In addition, due to the non-tamperability and
upgrading particularity of smart contracts, they are actually only a series of deterministic
and complex results fed back based on different inputs. Not only that, the code of smart
contracts is rigid and invariable, and lacks necessary flexibility in practical applications.
Therefore, current smart contracts are not smart, whose processing is deterministic.
AI brings good news to the relatively rough smart contract technology in blockchain
and helps to realize contract intelligence. First, AI combined with smart contracts can
quantitatively deal with problems in specific fields and make smart contracts own certain
prediction and analysis capabilities. For example, in the application of insurance anti-
fraud, the risk control model is built based on AI technology, the anti-fraud prediction
is carried out through the data combination of different telephone numbers of operators,
and the corresponding processing is carried out according to the rules of the smart
contract. AI-based smart contracts can handle financial risks that can’t be predicted by
human beings, and it has advantages over mankind in credit rating and risk pricing.
Second, the intervention of artificial intelligence makes it have the ability of bionic
thinking evolution. As far as the smart contract itself is concerned, through the AI
engine, under the guidance of the template and wizard program of the graphical interface,
the user input can be converted into complex smart contract code, that is, the “Smart
Protocol” that conforms to the user and the business scenario can be generated. Third, AI
constantly forms public computing power through learning and application practice. Of
course, the deep integration of AI and smart contracts needs to overcome both legal and
 Exploration on the Fusion of Artificial Intelligence 163

technical difficulties. Although some relatively simple contracts can usually automate the
performance, human intervention may be required to resolve disputes for more complex
contracts.

5 Conclusion
In this paper, we analyzed the advantages and disadvantages of AI and blockchain from
different levels of data, algorithm and computing power etc., and explore the necessity
and benefits to fuse AI and blockchain and make them complement each other. Through
extensive analysis, it can be seen that blockchain and AI can mutually empower each
other and it is a good idea to integrate them.

References
1. Marwala, T., & Xing, B. (2018). Blockchain and artificial intelligence. arXiv preprint arXiv:
1802.04451.
2. Norvig, P. Russel, and S. Artificial Intelligence. “A modern approach.” Prentice Hall Upper
Saddle River, NJ, USA: Rani, M., Nayak, R., & Vyas, OP (2015).
3. Wikipedia. Blockchain. Online: [Link]
4. Dinh, T. N., & Thai, M. T. (2018). AI and blockchain: A disruptive integration. Computer,
51(9), 48-53.
5. Aniello, L., Baldoni, R., Gaetani, E., et al. (2017, September). A prototype evaluation of a
tamper-resistant high performance blockchain-based transaction log for a distributed database.
In 2017 13th European Dependable Computing Conference (EDCC) (pp. 151–154). IEEE.
6. OpenAI. AI and Compute. Online: [Link]
7. Chang, Z. (2018). Video Replication and Access over Fog-based Architecture.
8. Salah, K., Rehman, M. H. U., Nizamuddin, N., & Al-Fuqaha, A. (2019). Blockchain for AI:
Review and open research challenges. IEEE Access, 7, 10127-10149.
9. Zou, W., Lo, D., Kochhar, P. S., Le, X. B. D., Xia, X., Feng, Y., & Xu, B. (2019). Smart contract
development: Challenges and opportunities. IEEE Transactions on Software Engineering,
47(10), 2084-2106.
10. Rahim, S. M., Mohamad, Z. Z., Bakar, J. A., Mohsin, F. H., & Isa, N. M. (2018). Artificial
intelligence, smart contract and islamic finance. Asian Social Science, 14(2), 145.

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 International License ([Link]
which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.