status

Artificial Intelligence and Cardano

Published 19.5.2023

ChatGBT is historically the fastest-adopted application on the Internet. AI is experiencing hype and many predict that this technology will change the world. Could something useful come out of combining AI and decentralization? Can Cardano be of any use for AI services? Or is it the other way around? Let's think about it.

TLDR

We see the biggest use of AI in assisting in programming smart contracts. Other uses include on-chain data analysis, information authentication, Oracles, etc.

AI as an assistant for smart contracts

AI can help developers write smart contracts. AI has enormous potential to replace or greatly supplement the work of programmers. In the future, programmers could be in the position of requesters. AI will be able to program a smart contract according to the specifications and verify its functionality.

The input provider may be experienced programmers, but one day also be ordinary users without IT experience. AI can write a smart contract tailored to their needs and deploy it straight away.

AI can be part of the user interface and will be able to explain to users in understandable human language what the current status of the smart contract is and what user input is pending (for example, sending tokens to a specific address).

Using conventional DeFi services is usually easy and straightforward for users. For complex contracts, AI can be a useful assistant to make the technology accessible to newcomers. A beginner who knows how to use AI well can ask if the smart contract he plans to use is safe.

Cardano is sometimes criticized for being difficult to write Plutus scripts because developers must know Haskell. AI can help write a secure smart contract in Haskell and additionally verify its reliability. Of course, the quality of a smart contract should be judged by a Haskell expert and not just AI. However, AI can speed up development significantly.

AI for on-chain analysis

Only experienced IT experts are able to perform on-chain analysis, as they know how to pull data from the blockchain and extract useful information from it. Linking AI to the blockchain will allow sophisticated questions to be asked by ordinary users. They can perform common queries, such as how many tokens are at a particular blockchain address, but also more complex queries, such as finding a token's history deep into the past.

AI algorithms can analyze the metadata, content, and transaction history of NFTs and provide evidence of their authenticity and origin. AI can assist with detecting fraud such as wash trading. AI can predict trends based on on-chain analysis.

What analyst firms are doing today will be available to everyone for free. Analytical firms may deliberately distort the results of the analysis for various reasons. AI will be unbiased and provide true results to specific queries.

Analytics firms focus on Bitcoin and do not have the capacity to produce reports for other blockchains. Of course, it is possible to create a dashboard with some charts without AI. However, AI can interpret the data in an understandable way and create graphs. It can also predict trends, detect abnormalities, etc.

How to authenticate information?

One of the big problems today is deep fake and the huge amount of information. Everyone has probably seen some AI-generated photos that looked like the real thing. Sooner or later it will be common to produce video content that is indistinguishable from the originals. The time is coming when we won't be able to trust everything we see on the internet. We are wary of textual information and know that much of it is false. In the near future, the same will be true for audio and video content.

People will be forced to verify the origin of information more thoroughly. With a decentralized identity (DID), it is possible to create a trusted link between an individual and their digital activities. When it comes to information, DIDs can serve as a means of proving the author, origin, and authenticity of the information, which can reduce the risk of false information and allow users to decide the credibility of the source.

I imagine that on Web3, users will choose people they trust and demand that they digitally sign the content they produce. Users will then verify the authenticity of the information every time before consuming it. Authentication can take place in the background. In this way, it will be possible to create an effective spam and fake news filter.

Information that is stored on the blockchain (not the entire content, but just a hash of the information or more information together) has the advantage of transparency and immutability. The public nature of the blockchain allows anyone to examine and verify the authenticity and origin of information through the records on the blockchain. In this way, transparency can be increased and the risk of falsification of information can be reduced.

There may be a version of AI that verifies the sources of information (digital signatures of the authors who provide the information) and provides results only on that basis. If there is any inaccuracy in the result, it will be very easy for the AI to track down the source of the inaccuracy and adjust its behavior.

In the flood of information and attempts to spoof false information, it will become increasingly important to verify the origin of information in a credible way. It can be laborious for humans but very simple for AI.

DAO as an opportunity to co-own and co-determine

Tokens on the blockchain have the advantage of being global as well as some AI services. If the team of an AI service wants to put decision-making power in the hands of users, they can do it through tokens on Cardano.

It doesn't matter if the tokens have market value or how the team distributes them. What matters is that the number of tokens can be capped. People can use tokens to vote on anything related to the service.

I can even imagine that the AI itself will evaluate the voting results and adjust its model. This is a bit of science fiction, but we may see it in the future. The advantage of using blockchain is that a third party cannot influence the vote in any way by manipulating the server.

Tokens or DIDs can be used for various forms of voting, for example on how many people believe a given piece of information or on who is or is not a trustworthy source of information. Consumers today have too many ways to interact with each other in a trustworthy way. Debates on social media can be inappropriately moderated. Voting by users in a transparent and immutable way may be much more important in the future than we currently think. Tokens (and NFTs) are a great tool for connecting community members.

AI as Oracles

Some smart contracts need to interact with the physical world and it is necessary to ensure the trust and reliability of this connection. The solution already exists, however, AI can greatly improve reliability.

Oracles usually work by taking information from multiple sources that are somehow economically incentivized not to cheat. Information from multiple sources is aggregated and passed to the smart contract.

AI can serve as one of the information inputs or as a control mechanism for the operation of Oracle. AI can be relatively reliable in detecting fraud attempts, abnormalities, trends, etc. If the inputs from Oracles are suspiciously inaccurate, AI can flag the problem.

We expect to be able to create an AI service that is autonomous. It might be a good idea to build a distributed network for AI and think about some form of decentralization.

AI and NFT

AI can generate a huge amount of digital content. Some works can be original and in high demand. In the flood of data, it can be useful to track down an original (first-of-its-kind) work. NFTs can serve this purpose very well, as by writing it to the blockchain, the time of creation and the author will be clearly defined.

AI can deliver an interactive and personalized experience for NFTs owners. For example, AI can be used to create interactive virtual worlds in which NFTs owners can interact with digital characters or environments. In this way, NFTs can offer dynamic and unique experiences that can change according to the owners' preferences and actions. NFTs can thus gain market value by offering a unique experience only available to the owner of a particular NFT.

Conclusion

AI is already being used today to analyze the reliability and security of smart contracts. Some audit firms take advantage of this. Using AI to analyze on-chain data can be relatively easy, and we can imagine something like this coming soon. For the other use cases described, connecting AI and Cardano (or other blockchains) may be more complicated, but we will probably see it sooner or later.

Technology is like Lego and innovative services are created by interconnecting them. Both AI and blockchain technologies will continue to evolve, so we can expect much more interesting use cases. The best is probably yet to come.

Featured:

Related articles

Did you enjoy this article? Other great articles by the same author