Putting data on blockchain doesn’t mean it’s correct

Until recently, the use of blockchain in elections was perceived as nothing more than an experiment. However, during the recent United States presidential election, some tried to turn the public’s perception of the possibilities of blockchain technology. For example, the Associated Press, one of the largest U.S. media outlets, published the election results on the Ethereum and EOS blockchains. 

Does this results call, however, suggest that the time to use blockchain in elections has come, and does it make sense to use the technology if the information source is centralized?

Being on a blockchain doesn’t make data trustworthy

Criticizing the AP as a data source may seem strange, given that it has been calling U.S. presidential elections since 1848. Perhaps, its nearly 200-year reputation was a reason why two decentralized platforms, YieldWars and Polymarket, chose to use AP data as an oracle for the 2020 elections prediction market. A YieldWars spokesperson called the AP “arguably the most trusted news outlet in the world.” Still, the word “arguably” suggests that data validity might still be in question.

And many people have indeed done just that — considered the AP an insufficiently reliable oracle for use in tandem with trustless blockchains. Some Twitter users met the news with both skepticism and outrage.

Commenting on the validity of such a data source, Juan Aja Aguinaco, co-founder of Shyft Network — a public attestation network for assigning context, trust and validation to data — noted an interesting trend: In some instances, particular outlets showed somewhat different results from one another. Moreover, according to him, the press is not responsible for making that kind of call, as “They do so because it’s what drives readership and ratings, but it’s not up to them to determine the winners of the elections.” He further assumed that the issues raised against the AP being considered an oracle might be valid, with a caveat:

“IF the purpose of using AP, or any other unofficial source of information, as an oracle for a prediction market is fair game as long as the participants fully understand what that means and the fact that AP can say one thing, but until the legal processes are over, the official results are undetermined.”

Still, the risk of any data fraud is minimal with media outlets such as the Associated Press, according to Thomas Stubbings, chairman of the Austrian Cyber Security Platform. He told Cointelegraph: “Probably few people would argue that AP is a more reliable source than, let’s say Breitbart News. Therefore, the reliability and trustfulness of a source like AP can be considered as given.”

By feeding bad data, the AP would effectively destroy its 100-year-old reputation as an unbiased reporter of elections, according to Artem Kalikhov, chief product officer of Waves Enterprise — the company whose technology was recently tested during Russia’s elections. He opined in an interview with Cointelegraph: “Since the data is cryptographically signed, the oracle node can’t manipulate it, only AP could taper with it, which is unlikely to happen.”

But what if the media is attacked by fraudsters? Although there is a possibility that a media outlet could be hacked and fake news spread, this would be noticed rather quickly. Stubbings said:

“Hacks which have a public impact are noticed very fast. As soon as there is any reasonable doubt, cybersecurity and forensic experts would jump in and examine the situation. And if there has been a hack or fraud — it will be found. Therefore, the possibility that an acknowledged medium can be hacked and fake information can be spread over a long time is absolutely impossible.”

He also suggested that a centralized media source could be more reliable than social media, which — ironically — seems to be more decentralized. “If it’s possible to centrally control such a media (like Facebook), it is possible to manipulate decentralized opinions from a central position,” said Stubbings, who added that this is what happened with Cambridge Analytica in 2016, where voters were centrally manipulated.

At the same time, Stubbings noted that any source is only as valid as the trust that is associated with it. The question “When is a source trustworthy?” is much more difficult to answer. Does this mean that the reputation of a trusted news source doesn’t guarantee that it’s actually trustworthy?

Decentralized oracles are not a solution

As it turns out, even if centralized, a trusted data source can supply information about elections to a blockchain. From that moment onward, the data cannot be deleted or changed. However, the question of how blockchain can verify the authenticity of the information remains open.

The problem is that today, smart contracts are not able to check whether the source of real-world information is reliable and complete. All that a smart contract can do is ensure the fulfillment of the prescribed conditions — for example, launching the function of replacing the president’s name on a platform after receiving information of their victory.

The good news is that there is a technology that can verify information, unlike a smart contract, and transfer it to a blockchain. These are trustless information providers — or oracles, as they are called in the blockchain space. However, not every information provider can be a true oracle. The oracle must be able to check the validity of the data — and, therefore, of the source of the information itself — and provide data on a wide range of events from the real world. Thus, having a reliable source of information is essential for the oracle to be reliable and complete.

In a conversation with Cointelegraph, Alice Corsini, chief operating officer at Provable Things — a platform that develops decentralized solutions, including oracles — agreed that when it comes to sensitive operations like political elections, it’s key for anyone to be able to verify the authenticity of data managed by oracles: “On this extent, oracles can adopt security technologies such as Trusted Computing for enabling data-authenticity verification and making the process transparent.”

Today, there are two main approaches to achieving the reliability of oracles. The first is oracle consensus, through which information is verified by several independent validators at once. In the second approach, the user themself chooses the source of information on the internet. Such a solution, for example, is offered by Provable Things, where TLSNotary proofs are used for proving the correct operation of the oracle. TLSNotary proofs provide cryptographic evidence that the data received from the selected source has been transferred to the smart contract unchanged.

Nevertheless, the problem of the reliability of the source itself remains unresolved. While both approaches guarantee the transfer of data from the source to the contract, they do not guarantee the integrity of the source, even if the oracle validators themselves chose it.

Speaking on the use of data published by the AP, Waves Enterprise’s Kalikhov suggested that although blockchain is already being used in national elections, this specific project does not bring real blockchain-based value to the voting process, as it’s only about fixing results in an immutable environment: “In case of oracle approach we still rely on traditional methods of vote collection and keeping vote secrecy before data gets to blockchain.”

More means better?

Some suggest that using multiple data sources and oracles together provides the best results in terms of the reliability and trustworthiness of the voting process. This means that using multiple media sources instead of just the AP might bring more trust to the process — even better if they are both local and foreign, and include social media.

An anonymous co-founder of YieldWars previously told Cointelegraph that future elections and prediction markets will be able to offer a more robust collection of oracles: “I envision there being multiple oracles like AP and I predict in the next election that we will see that. Having a number of trusted oracles settling markets should settle virtually all disputes.”

Kylin Network, a provider of decentralized oracles that recently received a Web 3.0 grant for building data infrastructure, offered to solve the problem of trustworthy data sources by collecting information about a particular event from indirect sources. So, the greater the number of these sources, the better. Dylan Dewdney, CEO of the platform, explained to Cointelegraph:

“So, to determine the election result, posts on social networks with the appropriate tags and date, the number of mentions of the presidential candidate on the Internet, publications in the media, etc. can be taken into account at the same time.”

Dewdney also noted that oracles must process large amounts of data simultaneously to ensure correct results. The best way to maintain that performance, according to him, is to make sure oracles have a stake in the game against a challenge or arbitration nodes.

This way, decentralized application developers can use such platforms to provide a validated premium data feed of their results of calls and validate all the API feeds to the chain. It’s in the interest of data providers to distribute accurate information because if it is challenged, they stand to lose the money they stake, as Dewdney added. “So, as a premium data feeder that has undertaken a validation process that is both decentralized and apolitical, the data I can provide — in this case, election results — becomes very valuable, and access to it, very valuable.”

In part, the experience of rewarding validators for providing information is used in the prediction markets. For example, the platform Augur uses the “wisdom of the crowd” principle to predict future events. Users predict the possible outcomes of these events by buying shares of the reward for correctly guessing the results. This approach results in economic motivation for the participants to ensure a correct prediction, and in the event that they are incorrect, they lose their stake. The forecast in this case is the weighted average of the expectations of all users.

The use of prediction markets greatly increases the completeness of the information provided, as anything can be predicted — if there are enough stakers — and reliability is provided by the economic motivation of the participants.

Has the time for blockchain in voting come?

Ultimately, the mere fact that the Associated Press interacted with blockchain to record voting results is not direct evidence that blockchain’s time in elections has come. Ashley Pope, co-founder of Fortis Block — a company that provides solutions for secure blockchain voting and digital elections for government, enterprise and nonprofits — claimed that instead, the news has shown the limitations and pain points of the current voting system:

“A large part of election processes worldwide are completed manually using a combination of paper/pencil/pen, and in some cases software. Voting is by and large stuck in the 1850’s. We bank online, pay taxes online and go to the doctor online yet voting is still completed manually. ”

Although the use of blockchain can technologically make elections transparent and reliable for voters, the problem of trust in authorities and the media, psychologically, may remain the same, according to Aguinaco: “Most people are distrustful of politicians and of the processes that get them elected. We could be using a 99% secure system and there would still be conspiracy theories, unrest, etc.”

In general, the use of blockchain in voting can have a positive effect on the electoral process. However, the transition to decentralized voting is not yet possible due to the laboriousness of organizing the process and its complexity for voters. It might, however, be more realistic in the short term to use decentralized oracles to validate votes. Although existing solutions provide a sufficiently reliable transfer of this information, the underlying issue of its original reliability still remains unsolved.