Launched in November 2022, ChatGPT has become possibly the fastest-growing software platform in history, reaching a million daily users in five days and 100 million monthly users within two months. Researchers believe monthly users now surpass 1 billion.
Add to that the users of Google’s Bard, Microsoft’s Bing, and the big generative art platforms and you can see how Bloomberg estimates that Generative Artificial Intelligence (AI) will generate annual revenue of $1.3 trillion within a decade. Given that 2022’s AI revenue was $40 billion, that’s a compound annual growth rate of 42 percent. That represents a major economic impact in hardware, software, marketing, infrastructure, and more.
But what will be the impact for crypto and crypto trading? Or DeFi?
Let’s take a look at some of the AI use cases founders and companies are employing right now or thinking about for the future in the crypto/DeFi world.
Any time a new and exciting technology emerges, one of the first things users ask is: “How can I make money with it?” In that vein, artificial intelligence and crypto trading are a match made in heaven. Crypto markets generate enormous amounts of data, and nothing is better at digesting and analyzing huge data sets than AI. Gathering the data sets necessary to perform various digital asset analyses is incredibly difficult and requires a massive data infrastructure, but once you have the data, AI can be extremely useful.
Here are some of the benefits AI offers crypto trading:
Investment robo-advisors have been a dream of retail investment firms for years. But for years they haven’t really delivered. Now, with Generative AI working off Large Language Models (LLMs), AI investment advisors’ time may have just arrived. Their ability to seemingly converse with users, as well as store and make sense of their risk tolerance and investment goals, make them easy to use and likely less biased than human advisors. If trained properly, they can also stay more up-to-date with the rapidly changing world of crypto assets than humans.
Both Morgan Stanley and JP Morgan Chase have already announced that they’re creating chatbot advisors, although Google and OpenAI advise against using their chatbots for financial advice. In the crypto markets, ATPbot has been called the “ChatGPT of crypto trading.” Besides pure investment advice, investment advice AI’s can also provide customer onboarding, financial education, personalized notifications, customer service.
Many crypto traders use AI to create and inform rules-based algorithmic trading bots that execute trades automatically based on predetermined rules and indicators. Sophisticated versions of these bots not only respond to real-time market activity and spot trends as they are forming, but can also learn from their own activity, adjusting and improving their strategy on the fly.
Besides being more efficient and faster, these bots of course eliminate human error, bias, and strategy drift. Some crypto bot creators also program their bots to automatically provide trading indicators, portfolio management, and backtesting.
One highly-regarded crypto trading bot creator is 3commas. Fully featured, it offers a complete suite of manual or fully automated bots and trading strategies as well as risk and portfolio management solutions. The platform allows users to deploy automated trading bots to execute non-stop cryptocurrency trading, minimize risks, and shop the marketplace for bot presets.
A well-regarded creator of crypto trading bots for beginners is Cryptohopper, which bills itself as “The world’s most customizable trading bot.” It claims to make two million trades per month. Ruud Feltcamp, one of the two brothers who founded the company, says “People ask me a lot, ‘Hey Ruud, I don’t trade crypto, but I want to use Cryptohopper, what should I do?’ My standard answer is, just start trading crypto first, and then see if automated trading is something for you.”
Retail investors are not the only ones seeking out the help of AI. AI can also be very useful for institutional traders working to develop and backtest algorithmic traders strategies.
Most successful trading strategies depend on the recognition of patterns within historical and current data series. If there’s anything AI can do well, it’s analyzing enormous amounts of data to find patterns. Via this analysis, it can help traders formulate and backtest trading ideas. It can even come up with its own ideas when prompted with various parameters by the user. AI models can analyze large data sets that include price history, transactions, order books, indicators, and other variables to backtest traders’ analysis.
There is a lot of fear among financial professionals about being left behind in the AI arms race, but there are also reasonable questions about AI’s reliability for trading. Trading isn’t social media, and it’s not meme generation; 90% accuracy isn’t acceptable. You can ask ChatGPT to build you a trading algorithm, and its output code looks good, but if you’re not a coder how do you know? Or how do you modify that code?
If you want to change the strategy for any reason, either the code initially didn't work or you just find it's not that good of a strategy and you want to change it, your options are to ask ChatGPT again, from scratch, and maybe massage the results via a number of subsequent queries. You’re on a one-way street. All it's doing is churning out code based on your request, and when you make a new request, or try to modify it, it's essentially like you're playing a new game. It’s a complete reset.
One company that is answering this challenge is 4Thought Technologies. 4Thought, on a grant from the National Science Foundation, is building a professional-grade no-code platform for AI-enhanced automated trading & analytics. A platform that enables serious professional traders to create their own algo trading strategies without the need for expensive coders.
Their platform operates with its own visual drag-and-drop language of pre-built blocks of code connected to accurate, validated data sources. This solves the data reliability problem. One potential source of reliable data is Amberdata, which provides the clean, verified, and consistent digital asset historical and live data that crypto traders need for trading, analytics, and backtesting.
A second AI algo trading problem is: if your algorithm isn’t giving you the results, how do you modify your algorithm without starting from scratch? AI-generated code is only so useful if you’re not a coder. “We assume that most traders aren’t coders,” says 4Thought CEO, Richie Sater. “So we let you do that in flowchart form, drag and drop. As far as we know, this is the first time in the financial space that users are able to manipulate the output of the generative AI in an easy-to-use way.”
Machine learning algorithms enable traders and investors to anticipate potential risks and proactively implement preventive measures. AI aids traders in enhancing risk management by evaluating data and identifying possible pitfalls, thereby helping them to make better decisions regarding positions and position sizing. In the future, AI-driven risk management chatbots offering round-the-clock customer support are likely to become widely available.
AI can continuously monitor cryptocurrency markets and track price movements across multiple exchanges. It can detect sudden price fluctuations or anomalies that may indicate potential arbitrage opportunities.
AI can also identify and exploit crypto arbitrage opportunities by leveraging its ability to analyze vast amounts of data and detect patterns that may go unnoticed by humans. AI algorithms can analyze real-time data from multiple cryptocurrency exchanges simultaneously. By monitoring price discrepancies and order book imbalances across different exchanges, AI can identify potential arbitrage opportunities.
Additionally, AI-powered trading systems can execute trades at high speeds, taking advantage of fleeting price disparities before they correct. AI can process large amounts of data in fractions of a second, enabling traders to capitalize on arbitrage opportunities that require swift execution. It can also instantly factor all the cost elements of an arb, considering transaction fees, withdrawal limits, and current market volatility, and likely transaction times. Such instant accounting is often the difference between a profitable execution and a loss.
Most interestingly, AI can use historical market data and machine learning algorithms to anticipate potential price disparities before they appear.
Another way that AI is transforming the crypto-trading is via sentiment analysis. It uses natural language processing (NLP) algorithms to sift through many online sources, such as social media interactions, stock message boards, blockchain data, and more, finding and evaluating information pertinent to specific cryptos or markets. The insights gathered from this analysis are then used by traders and investors as part of their investment decision-making process.
A firm currently offering AI-driven crypto sentiment analysis is Stockgeist, an interactive platform that provides real time crypto market sentiment analysis for over 350 different assets, derived from textual messages across social media platforms. StockGeist is powered by AI deep learning-based solutions.
Predictive analytics leverages the power of AI to enable traders to anticipate future market trends. This is achieved by scrutinizing historical performance data with the help of complex mathematical models such as regression analysis and time series forecasting.
The projections generated through these models can provide valuable insights into potential price fluctuations across different markets, thus aiding in making winning investment decisions. Additionally, AI can improve its predictive skill over time via machine learning.
Once a particular trading strategy has been determined, either by AI or a human, the AI can use smart contracts to execute trades on-chain. Trades will be made automatically when specific criteria are met. Not only does this increase efficiency, but it eliminates human error. Presumably, it also manages risk, since the predetermined parameters will be in part based on individual trader risk tolerance.
AI can assist in finding liquidity in crypto markets. In crypto markets, liquidity is critical as it influences the stability of the asset's price and the speed at which it can be converted to cash or another crypto. AI, coupled with machine learning, can analyze large amounts of data from various exchanges in real-time, helping traders and investors identify liquidity opportunities. For instance, an AI system can be programmed to track and analyze the order books of various exchanges simultaneously.
This real-time analysis can then be used to identify where liquidity exists at any given moment. Furthermore, AI algorithms can predict liquidity changes based on historical data and trends, and can also be used to create predictive models that forecast market movements, which can help in planning entries and exits, thereby potentially optimizing trade execution and reducing the impact of your larger trades on the market.
Fluid is a low latency crypto liquidity aggregator that uses AI-driven prediction models to make trade execution smoother and more efficient, with accurate pricing.
Where there are crypto projects there are crypto tokens. Accordingly, for all the new AI enterprises being birthed by LLMs and other AI models, there are many new AI tokens (as well as some for older projects).
As with most tokens, these AI tokens typically confer governance rights or rewards on the holders, but many also have uses more specific to their platform. While the market capitalization of AI crypto is still fairly small, less than $2 billion, many of the projects they back are interesting.
Here are a few of them:
Decentralized finance (DeFi) refers to the growing number of financial applications and services offered on decentralized, transparent blockchain networks. Crypto loans, yield harvesting, and peer-to-peer lending are all examples of DeFi.
The growth of DeFi over the last eight years, made possible by the launch of the Ethereum blockchain, has been extremely rapid. Combine that with the explosive growth of AI and you have the potential for future products and services that are truly disruptive.
What might these be?
Let’s take a look at how innovators are combining DeFi and AI right now.
AI-generated algorithms can automate trading processes and execute trades on decentralized exchanges. Also, as mentioned above, AI tools can uncover trends, arbitrage opportunities, and a variety of profitable strategies and tactics if fed accurate historical crypto data.
AI can provide personalized investment advice as well as automating portfolio management for DeFi traders. Investment “advicebots” can be customized to the platform’s AI system considers factors such as an individual’s risk tolerance, investment goals, and historical performance to make personalized recommendations, helping users to maximize returns and minimize risks.
Much of trading in any financial market involves the analysis of large historical trade price datasets, seeking out patterns and indicators that might inform future price direction. AI can, of course, analyze these datasets better and much faster than humans. Additionally, over time, AI can refine its predictive capabilities. Amberdata can provide these clean historical data sets.
AI can be utilized in order to improve the efficiency of smart contracts. Right now, various lending protocols are using AI to monitor and adjust individuals’ collateral levels. Furthermore, AI can assess all of borrowers’ past and current activity to create sophisticated credit scoring to determine individualized interest rates and loan terms. This is, of course, a benefit to both borrowers and lenders.
Miami tech startup QUASH.ai recently raised $3.7 million to expand into Latin America. Its AI engine enables lenders to assess and classify loan applicants more accurately, increasing loan volume and profitability, and decreasing defaults.
AI agents are software bots that can do specific tasks all by themselves using LLMs, without human guidance. Integrating AI into blockchains enhances the decision-making or predictive capabilities of smart contracts, utilizing dynamic on-chain data. This creates the possibility of many new capabilities. As these AI agents get more advanced, they'll get better at doing specific jobs. In all likelihood, this will lead to a marketplace for AI agents, where AI agents can hire each other on the fly to do certain tasks, paying each other with cryptocurrency. They could also tap into other resources like computing power and storage. A brave new world, indeed.
One company creating infrastructure for autonomous AI agents is Fetch.ai, which is building an open-access decentralized machine learning network for smart infrastructure. Using Fetch technology it is possible to deploy AI agents to create intelligent connections between each other and real-world systems and devices. Some of the agent capabilities Fetch supports are swap support, ETH fund management, APY monitoring and automatic withdrawals, strategies for pool deposits and withdrawals, and portfolio management.
DeFi protocols are vulnerable to various types of attacks and risks. AI can be used to predict such risks, increasing the safety of protocol investor funds.
DeFi protocols increasingly must adhere to a number of local regulations in the jurisdictions in which they have users. The most common of these are Know Your Customer and Anti-Money Laundering regs. These processes can be onerous; AI can streamline and automate them.
DeFi and crypto’s inherent anonymity has, since their appearance, enticed bad actors wishing to commit fraud of one kind or another. AI can help detect this behavior by analyzing enormous datasets to unearth trends of behavior. Currently, AI is being used to help spot wash trading volume, which inflates coin and exchange values, and can inflate liquidity. Used by governments and law enforcement around the world, Chainalysis is the crypto industry standard for detecting fraud, money-laundering, market manipulation, and other fraudulent behavior.
As Decentralized Autonomous Organizations (DAOs) are becoming more popular as business and project entities, the question arises as to how AI can help. One obvious way is to manage DAO treasuries to buffer them from risk and maximize returns on free cash when advisable.
Smart contracts are what make DeFi possible, but developers have struggled with the fact that once a protocol becomes valuable, floods of hackers and other malicious actors scour the code of the contracts looking for loopholes and bugs to exploit. Now, however, AI’s can audit code, and they have proven to be very good at it. AI can also monitor contracts in real time for anomalous behavior, enabling them to respond quickly.
Empowering smart contracts with the ability to interpret AI and machine learning models greatly expands the potential of what can be engineered on-chain. It can improve existing projects by making them more efficient and introducing new features. Until recently, the integration of AI into smart contracts has been challenging due to the extensive computational resources required to deploy the model on-chain.
However, now a solution to this problem resides in Zero-Knowledge Proofs. One can use a Zero-Knowledge Proof to verify the correct off-chain execution of a model and post this proof on-chain, thereby adding the intelligence generated by AI models into smart contracts. Implementing AI on-chain could also allow smart contracts to make decisions based on real-time on-chain data rather than static rules. This could lead to more sophisticated smart contracts.
Artificial Intelligence has revolutionized crypto and DeFi by providing inexpensive tools to automatically analyze and act upon the vast datasets that blockchains and crypto trading generate. The benefits of using AI here now and in the future cannot be overstated. AI represents a major step forward for both individual consumers, borrowers, investors, and traders as well as financial institutions looking to stay ahead of the curve. Whoever you are, if you transact or trade, AI can be one of your best new tools.
Your other best new tool can be Amberdata, which will enable you to get the data you need to feed your AI. Generating or obtaining the large, clean, and accurate datasets you need is very difficult and requires enormous data infrastructure. Amberdata’s unified API and data services provide you a single integration point for obtaining a comprehensive view of the entire cryptoeconomy.
We deliver comprehensive digital asset data and insights into blockchain networks, crypto markets, and decentralized finance, empowering financial institutions with critical market or DeFi data for research, trading and analytics, risk management, derivatives analytics and compliance. To learn more about Amberdata, please contact us to book a demo, hear about our products that can help your business, or receive pricing information.