Page 237 - Artificial Intelligence for the Internet of Everything

218 Artificial Intelligence for the Internet of Everything

through a complex inefficient multitiered system where consumers have lit-
tle control in the type of energy they consume. Blockchain is defined as a
distributed database or digital ledger that records transactions of value using
a cryptographic signature that is inherently resistant to modification
(Tapscott, 2016). Combining blockchain-based smart contracts with
machine learning and AI algorithms presents an opportunity to realize the
goals of DAEOs. DAEO will provide a platform for a more distributed
autonomous system that helps increase the speed, scale, security, and auton-
omy of complex, distributed Internet of Things (IoT) environments, such as
the US power grid. The need for third parties in executing a complex energy
transaction may be reduced when an autonomous smart contract can exe-
cute and exchange value and services via an autonomous agent or even a
DAEO blockchain platform. Smart contract design and DAEO can also help
inform machine-learning algorithms. Like many proof of work (PoW)
blockchain technologies, DAEO would not need to have a single user with
control. Blockchain platforms and associated smart contracts could help cre-
ate some control mechanisms to retain control of the DAEO to an auton-
omous AI agent. Blockchain technology could help secure the sensitive data
needed to inform the AI in a DAEO because the data would be cryptograph-
ically signed and stored securely in a distributed ledger, enabling access, data
provenance, and auditability while preserving privacy. This is especially
important if the AI will be informed with sensitive or personal identifiable
information.
Blockchain technology provides a more efficient ledger for developing,
securing, and improving AI and machine-learning algorithms
(Zwanenburg, 2018). AI technology continues to make rapid advances in
everything from translation to surgery to virtual assistants and self-driving
vehicles. But people also fear AI. Opponents of AI point to pedestrian deaths
caused by AI-enabled self-driving vehicles and studies that suggest it will
replace more jobs than it creates. Developing AI also requires access to
the use of large data sets that could potentially compromise people’s privacy.
In that respect, blockchain can help preserve the privacy of the data sets
needed for large neural networks by storing a cryptographic hash of the
metadata that can only be unlocked via a private key and permissioned access
to the blockchain. Accordingly, blockchain can help AI developers better
understand and track the prodigious data sets, increasing the auditability
and control of neural-network machine-learning functionality. The ability
to track and control large data sets is especially important to provide distrib-
uted autonomous energy organizations with reliable information because