ИИ как судья: Как искусственный интеллект может изменить рынок предсказаний Translation: AI as a Judge: How Artificial Intelligence Can Transform the Prediction Market

Artificial intelligence has the potential to serve as an integrated judge within blockchain-based prediction markets. This viewpoint was articulated by Andrew Hall, a professor of political economy at the Stanford Graduate School of Business.

He illustrated the challenge of «fair» dispute resolution using the example of the presidential elections in Venezuela.

Last year, contracts totaling over $6 million were placed on the outcome of the event. However, following the campaign, the market was left in confusion:

«Should the outcome of the prediction market contracts adhere to the ‘official’ information (the victory of Maduro) or to the ‘consensus of credible reports’ (the opposition’s victory)?» Hall questioned.

This is not an isolated instance, the expert pointed out. In another case, manipulations allegedly occurred regarding the territorial dispute over Ukraine.

Hall emphasizes the necessity of establishing a fair contract resolution system that people can trust. In such a scenario, prices can serve as meaningful signals for society.

Similar issues plague financial markets as well. For years, the International Swaps and Derivatives Association has tackled challenges related to the resolution of credit default swaps—agreements that trigger payments in the event of a company’s or country’s bankruptcy.

Decision-making committees vote on whether credit events have occurred, yet this process has faced criticism for its lack of transparency, potential conflicts of interest, and inconsistent outcomes.

«The fundamental issue remains the same: when large sums depend on determining what transpired in ambiguous situations, any resolution mechanism becomes a target for manipulation, and ambiguity becomes a potential point of contention,» Hall stated.

The expert listed several key attributes that any viable solution should possess:

Human committees may satisfy some of these criteria, but they are susceptible to manipulation and cannot maintain neutrality.

Hall proposes utilizing large language models as judges, with each model and prompt being recorded on the blockchain at the moment the contract is created.

The basic architecture is as follows:

This method addresses several major challenges:

On the downside, AI can make mistakes. It might misinterpret a news article or create a fabricated fact.

Manipulation is possible, but it becomes more complex to execute. Fraudsters could arrange for certain information to be published in major media outlets. While costly, this is feasible.

There is also the risk of attacks on the training data of LLMs, but this requires action long before the contract is signed.

AI-based solutions merely replace one set of problems with another, more manageable set. Hall believes platforms should experiment with different LLMs to gain experience.

As best practices emerge, the community must work towards standardizing combinations of AI programs. This will help concentrate liquidity, according to the author.

It’s worth noting that in January, analysts from a16z crypto predicted a growth in prediction markets and ZK-proof technologies.