Crypto Companies Leverage AI Despite Ongoing Challenges

Publikováno: 26.1.2024

Recent statistics show that the artificial intelligence (AI) market size is expected to reach over $3 billion this year. Therefore, it shouldn’t come as a surprise that a number of crypto-focused companies have started to incorporate AI into their products. Why companies are combining crypto with AI Jacqueline Burns-Koven, head of cyber threat intelligence for […]

The post Crypto Companies Leverage AI Despite Ongoing Challenges appeared first on Cryptonews.

Celý článek

Recent statistics show that the artificial intelligence (AI) market size is expected to reach over $3 billion this year. Therefore, it shouldn’t come as a surprise that a number of crypto-focused companies have started to incorporate AI into their products.

Why companies are combining crypto with AI


Jacqueline Burns-Koven, head of cyber threat intelligence for Chainalysis – a blockchain analysis firm – told Cryptonews that Chainalysis has started thinking about ways to use AI to make compliance, risk, investigations and growth products better for customers. “Like any business, we stand to benefit by utilizing AI to improve how we work across the business; by making it faster and more efficient,” Burns-Koven said.

Crypto tax software provider ZenLedger also recently announced a partnership with april – an AI-powered financial company – to use AI to simplify the tax filing process for users. Pat Larsen, co-founder and CEO of ZenLedger, told Cryptonews that ZenLedger’s new product will leverage april’s technology to route taxpayers through a single flow, combining federal and state, and then deciding which question to ask next. “This is in contrast to traditional tax filing software that asks questions of the user in the order the forms are completed, and then parses out federal and state forms into separate sections, often duplicating the same questions in each,” Larsen said.

Daniel Marcous, CTO and co-founder of april, told Cryptonews that AI has been instrumental to april’s ability to build a tax product covering many common tax scenarios, including income from crypto and digital assets. According to Marcous, april uses a process called “tax-to-code” in which large language models (LLMs) have been trained to read tax documents and then turn those into software code which is then reviewed and edited by a team of tax engineers.

AI is also helping power a number of decentralized finance (DeFi) use cases. Nick Emmons, co-founder and CEO of Upshot – an AI infrastructure company – told Cryptonews that Upshot is building a decentralized network where different AI models can learn from each other. According to Emmons, having models learn off each other will create a meta intelligence across an AI-powered network. In turn, this will make networks more performant and intelligent compared with individual models being used.

Emmons explained that Upshot’s AI model is powering a number of DeFi use cases. For example, he explained that AI can create efficiencies for price feeds for long-tail crypto assets, or digital assets that don’t often trade but exist in liquid settings. He said:

“AI becomes a useful tool for being able to produce more frequent price updates based on different information, not just an asset changing hands. This means that we can now start to bring a much larger universe of assets into the DeFi design space.”

To put this in perspective, Emmons explained that Upshot will soon introduce “watch perks” generated by AI-enabled watch feeds. He said:

“An individual watch is incapable of producing a real enough time feed to build a market around it. AI models can process a lot of information at once, so you can start to produce highly accurate and high frequency price feeds to turn digital assets into on-chain, tokenized representations. This will expand the universe of digital assets.”

Additionally, Emmons pointed out that AI-powered DeFi vaults are coming to fruition. A DeFi vault acts as a pool of funds with an auto-compounding strategy that manages and performs tasks based on predefined on-chain conditions. Yet Emmons noted that this is problematic given that most on-chain activity is limited when it comes to compute power. “As such, the yield a user can generate is limited,” he said.

In order to solve this problem, Emmons noted that AI models can be applied to make sense of information more efficiently. “AI can be used to codify strategies that can be brought on-chain in the form of vaults. This can then be used for market making and more.”

Although this use case is still in its infancy, RoboNet is an AI-powered DeFi protocol for long-tail and fungible asset markets. RoboNet is powered by Upshot and allows for the creation of on-chain vaults managed by machine learning models that generate yield through automated liquidity optimization strategies.

Source: Robonet

Challenges combing AI with crypto


While AI can help crypto products perform more efficiently, there are still a number of challenges to consider. For example, Emmons pointed out that when AI is leveraged for building DeFi protocols, the creators behind those models need to be trusted, otherwise a number of issues could occur. He said:

“Bias and manipulation can arise, which is why it’s important to reimagine the AI stack in decentralized form factors. Different models can keep other models in check to create less bias and a more transparent source of intelligence.”

Emmons explained that ZK proofs can also help verify machine learning models. “Upshot recently released a product like this where we verified the output of our flagship price prediction model inside a ZK circuit. This provides assurance and computational integrity for permissionless protocols.”

Marcous added that he believes generative AI working alongside tax experts and engineers mitigates risk since a human is involved. “At april, we conduct a rigorous testing process on the entirety of the product and have to pass tests with the Internal Revenue Service and state authorities before launching,” he said.

While these tactics may be helpful, the lack of regulations around the use of AI will likely present ongoing challenges. For instance, understanding whether or not AI is being applied for the best interest of users versus investors or the creators of machine learning models remains difficult to determine.

Due to this, certain countries have started to establish organizations to enforce AI regulations. For example, the president of the United Arab Emirates and ruler of Abu Dhabi, Sheikh Mohamed bin Zayed Al Nahyan, recently issued a law to establish the Artificial Intelligence and Advanced Technology Council (AIATC). An announcement from the Abu Dhabi government noted that, “the council will be responsible for developing and implementing policies and strategies related to research, infrastructure and investments in artificial intelligence and advanced technology in Abu Dhabi.”

United States Securities and Exchange Commission (SEC) Chair Gary Gensler also recently warned about the dangers that AI could pose to the traditional financial sector. Given this, more regulatory clarity around AI will likely be implemented in the U.S. in the future.

All of these developments are important, as Emmons believes that AI will eventually be incorporated into every critical function of society. In the meantime, he pointed out that the crypto sector will likely incorporate forms of AI that have already been implemented in traditional financial systems. He said:

“This is because crypto is a financial innovation, so this type of AI can be more conducive with financial applications. Also, classical types of machine learning models are more attractive and compatible with these verifiable form factors, so cryptographic tooling that can be built around those will be able to come online faster than generative AI models.”

The post Crypto Companies Leverage AI Despite Ongoing Challenges appeared first on Cryptonews.

Nahoru
Tento web používá k poskytování služeb a analýze návštěvnosti soubory cookie. Používáním tohoto webu s tímto souhlasíte. Další informace