How AI can and will be used in Decentralized Finance is a popular discussion at this time and some are wondering the scope of what it could do for the crypto space and how to achieve it. Autonio has been hard at work this past year on AI research with the help of Anton Kolonin, a blockchain and AI architect, one of SingularityNET’s senior AI researchers, and a team member of the Autonio Foundation. Now is a great time to ask him some questions about the work he’s been doing.
What are the skills one needs in order to work and create with AI?
- There are many paths one can take in this area, some are more mainstream and some less. The most mainstream one would be to learn Python programming, linear algebra, vector math, and statistics, then find some intern/junior work to start contributing to some real projects on the market relevant to those skills. Describing the many non-mainstream paths you could take in AI would take significantly more space and would be dependent on what specific AI you need. The review on the matter that I wrote some time ago is here and my next review on the subject will be published on the same site by the end of this year.
How would you describe the relationship between AI and blockchain?
- It’s a loose relationship regardless of the marketing hype around it. It is like what is the relationship between the engines and wheels - both are used in vehicles, although engines can be also used in boats and wheels can be used in rollerblades. Moreover, the blockchain as we know it is just the very specific, the most simple, the least efficient and the most expensive version of what is called a “distributed ledger.” More advanced versions of this concept are in active development in projects like TODA (https://www.toda.network/). I have some papers in Russian written on “AI and blockchain” written a while ago here and the latest update on “blockchain” matter is available in English here.
On a practical level, what are some of the ways that AI could enhance the DeFi space?
- AI is a way to automate anything, as it already automates conventional Financial Technology (FinTech), it can be used to automate DeFi operations as well. See more in our publication here and our YouTube video here.
What are some of the limitations that AI might run up against in making an impact in the industry?
- The nature of the crypto and DeFi market is substantially more chaotic, manipulative, and non-predictive than in the case of conventional FinTech. As a result, the application of AI encounters major complexities here and we still have to see to what extent they can be overcome.
Since DeFi is essentially built on a bunch of protocols, is there a possibility that AI could somehow hack into these?
- I think it’s safe to assume that all these protocols are attempted at being hacked continuously as we speak. Consider the “flash loan attack” for example. AI can just multiply the scale and impact of such hacks, so one would need another “anti-hack” AI to resist the “hacky” AI, similar to how it happens in conventional cybersecurity.
Furthermore, could the advancement of AI come at the expense of human livelihoods?
- Consider the fire. It helped humanity to survive during the glacial periods and helped them to get smarter by spending less resources to digest pre-cooked meals. On the other hand, it is killing thousands of people per year in building fires, while wildfires are destroying our environment. Then we have nuclear energy, the most clean and “green” power with the least carbon footprint but also the most dangerous, with the potential to end the whole of civilization at the same time. AI is kind of like that. It can do a lot of good, but just as well, can be used for bad. Finally, if you think hard at the scale of the history of the Universe, you might figure out that AI is one of the ways it could be possible to have descendants of biological humans in the really distant future - consider this perspective.
Have fun trading with Autonio!