Artificial Intelligence – an overview | ScienceDirect Topics

12.10 Conclusion and Future Research

AI blockchain enabled distributed autonomous energy organizations may help to increase the energy efficiency, cyber security, and resilience of the electricity infrastructure. These are timely goals as we modernize the US power grida complex system of systems that requires secure and reliable communications and a more trustworthy global supply chain. While blockchain, AI, and IoT are creating a buzz right now, many challenges remain to be overcome to realize the full potential of these innovative technological solutions. A lot of news and media coverage of blockchain today falsely suggests that it is a panacea for all that ails usclimate change, cyber security, and volatile financial systems. There is similar hysteria around AI, with articles suggesting that the robots are coming, and that AI will take all of our jobs. While these new technologies are disruptive in their own way and create some exciting new opportunities, many challenges remain. Several fundamental policy, regulatory, and scientific challenges exist before blockchain realizes its full disruptive potential.

Future research should continue to explore the challenges related to blockchain and distributed ledger technology. Applying AI blockchain to modernizing the electricity infrastructure also requires speed, agility, and affordable technology. AI-enhanced algorithms are expensive and often require prodigious data sets that must be broken down into a code that makes sense. However, a lot of noise (distracting data) is being collected and exchanged in the electricity infrastructure, making it difficult to identify cyber anomalies. When there is a lot of disparate data being exchanged at subzero-second speeds, it is difficult to determine the cause of the anomaly, such as a software glitch, cyber-attack, weather event, or hybrid cyber-physical event. It can be very difficult to determine what normal looks like and set the accurate baseline that is needed to detect anomalies. Developing an AI blockchainenhanced grid requires that the data be broken into observable patterns, which is very challenging from a cyber perspective when threats are complex, nonlinear, and evolving.

Applying blockchain to modernizing and securing the electricity infrastructure presents several cyber-security challenges that should be further examined in future research. For example, Ethereum-based smart contracts provide the ability for anyone to write electronic code that can be executed in a blockchain. If an energy producer or consumer agrees to buy or sell renewable energy from a neighbor for an agreed-upon price, it can be captured in a blockchain-based smart contract. AI could help to increase efficiency by automating the auction to include other bidders and sellers in a more efficient and dynamic waythis would require a lot more data and analysis to recognize the discernable patterns that inform the AI algorithm of the smart contracts performance. Increased automation, however, will also require that the code of the blockchain is more resilient to cyber-attacks. Previously, Ethereum was shown to have several vulnerabilities that may undermine the trustworthiness of this transaction mechanism. Vulnerabilities in the code have been exploited in at least three multimillion dollar cyber incidents. In June 2016 DAO was hackedits smart contract code was exploited, and approximately $50 million dollars were extracted. In July 2017 code in an Ethereum wallet was exploited to extract $30 million dollars of cryptocurrency. In January 2018 hackers stole roughly 58 billion yen ($532.6 million) from a Tokyo-based cryptocurrency exchange, Coincheck, Inc. The latter incident highlighted the need for increased security and regulatory protection for cryptocurrencies and other blockchain applications. The Coincheck hack appears to have exploited vulnerabilities in a hot wallet, which is a cryptocurrency wallet that is connected to the internet. In contrast, cold wallets, such as Trezor and Ledger Nano S, are cryptocurrency wallets that are stored offline.

Despite being a centralized currency, Coincheck was a cryptocurrency exchange with a single point of failure. However, the blockchain shared ledger of the account may potentially be able to tag and follow the stolen coins and identify any account that receives them (Fadilpai & Garlick, 2017). Storing prodigious data sets that are constantly growing in a blockchain can also create potential latency or bloat in the chain, requiring large amounts of memory. Requirements for Ethereum-based smart contracts have grown over time and the block takes a longer time to process. For time-sensitive energy transactions, this situation may create speed, scale, and cost issues if the smart contract is not designed properly. Certainly, future research is needed to develop, validate, and verify a more secure approach.

Finally, future research should examine the functional requirements and potential barriers for applying blockchain to make energy organizations more distributed, autonomous, and secure. For example, even if some intermediaries are replaced in the energy sector, a schedule and forecast still need to be submitted to the transmission system operator for the electricity infrastructure to be reliable. Another challenge is incorporating individual blockchain consumers into a balancing group and having them comply with market reliability and requirements as well as submit accurate demand forecasts to the network operator. Managing a balancing group is not a trivial task and this approach could potentially increase the costs of managing the blockchain. To avoid costly disruptions, blockchain autonomous data exchanges, such as demand forecasts from the consumer to the network operator, will need to be stress tested for security and reliability before being deployed at scale. In considering all of these innovative applications, as well as the many associated challenges, future research is needed to develop, validate, and verify AI blockchain enabled DAEOs.

Read the original here:
Artificial Intelligence - an overview | ScienceDirect Topics

Related Posts

Comments are closed.