- Types of crimes that use AI
- How AI is being used for cryptocurrency-related deception
- Token and AI scams
- State-sponsored hackers trying to integrate artificial intelligence
- Suggestions for protecting the crypto market from threats
A new study by Elliptic has revealed that AI-assisted crimes in the cryptocurrency ecosystem are still in their nascent stages. Stakeholders can prevent the spread of these activities with timely and measured responses. The data also indicates a spike in the number of tokens containing AI-related keywords such as GPT, Openai, and Bard. Approximately 4,500 of these tokens are present on the BNB smartchain.
Types of crimes that utilize AI
A study conducted by Elliptic outlines five typologies of AI-enabled crime that key stakeholders should consider to prevent escalation.
- Generative AI for crypto fraud, including the use and distribution of deepfakes and AI-generated materials to publicize crypto fraud.
- AI-related crypto fraud and market manipulation schemes, including the creation of ether-linked fraudulent crypto tokens, investment platforms and pyramid schemes, which are increasingly common.
- The use of large language models (LLMs) to combat cybercrime, including the use of artificial intelligence tools by hackers and hostile government agencies to detect vulnerabilities in code and develop exploits.
- Large-scale deployment of crypto fraud and disinformation using artificial intelligence tools.
- Expansion of illicit markets, including artificial intelligence-enabled expansion and the creation of illicit economies for goods and services, such as online darklists, the creation of blatant fakes or fake identities that bypass know-your-customer (KYC) verification in crypto services.
Artificial intelligence adds a new dimension to the crime problem by allowing fraudsters to impersonate famous individuals with greater ease using images, video and voice. Edited videos – or “deep fakes” – of famous personalities advertising investment scams have been targeted to resemble Ilon Musk, former Singaporean Prime Minister Lee Hsien Loong and Taiwanese presidents Tsai Ing-wen and Lai Ching-te. Advertising fakes are often posted on sites such as YouTube, Tik Tok and x.com, the latter of which is owned by Musk himself. In one particularly brazen use of “deepfake,” cybercriminals published a video of former Singaporean Prime Minister Lee Hsien Loong endorsing a cryptocurrency scam.
Fake screenshots of former Singapore Prime Minister Lee Hsien Loong (left) and Taiwan’s 7th President Tsai Ing-wen promoting cryptocurrency investment. Source: Lee Hsien Loong (Facebook) and Taiwan’s Central Bureau of Investigation, respectively.
How AI is being used for crypto-related deception
There are a number of aspects in which AI can be misused to make crypto fraud and scams more convincing:
- Hinting at the involvement of celebrities or officials: as the leaders in Singapore and Taiwan highlight, fakes can falsely claim that a project has legitimate or official backing, thereby legitimizing it among potential victims.
- Using artificial intelligence to optimize fraud profiles and communications. Some evidence suggests that these illicit operations are exploring improvements in artificial intelligence to make these processes more efficient.
- Deepfake fraud: similar to deepfake fraud impersonating celebrities or world leaders, a small number of high-profile cases have involved fraudsters impersonating high-level executives during online video conferences for the purpose of corporate espionage or authorizing large transactions.
- Creating fake marketing materials: artificial intelligence-generated images and videos can be used to create an aura of legitimacy for fraudulent websites by depicting purported employees, headquarters, office space, or other similar visual elements, creating the illusion that it is a real investment company without revealing the actual individuals involved or the location of the fraud.
Token and AI fraud
In many blockchains, creating a token requires little effort. Many fraudsters take advantage of this opportunity, often capitalizing on recent newsworthy events to drum up hype. Beyond the erroneous assumption that there is a commercial connection between a legitimate company and a crypto token, such attempts can lead to two main types of fraud:
- Criminals blow the hype to raise the price of their token and then sell their reserves at a significant profit, causing the price to fall again and leaving their victims with a completely worthless investment. In 2021, a token named after the popular Netflix series “The Squid Game” gained notoriety after scammers made several thousand dollars.
- Market manipulation schemes in which coordinated groups initiate sudden purchases and sales of tokens for profit.
Artificial intelligence is the object of hype in a number of recent token scams that have several variations of the term ‘GPT’ in their names, including ‘GPT4 Token’, ‘GPT Cryptographic Protocol’ and ‘GPT Coin’, among others. Some of these may indeed be well-intentioned, but some have been posted on amateur trading forums where scammers claim any official connection to ChatGPT or another supposedly legitimate artificial intelligence company.
Number of tokens named with AI-related keywords per blockchain
State-sponsored hackers are trying to integrate artificial intelligence
The report notes that artificial intelligence models such as ChatGPT, which can detect vulnerabilities in smart contracts, are now being used by hostile state-sponsored actors to commit cryptocurrency thefts. For example, North Korea-backed hackers are attempting to use artificial intelligence to enhance their capabilities to conduct large-scale heists.
“The United Nations has attributed more than 60 cryptocurrency thefts to hostile state actors based in North Korea, resulting in more than $3 billion worth of cryptocurrencies stolen between 2017 and 2023. Recent reports and future threat assessments suggest that North Korean groups are turning to artificial intelligence to expand their hacking capabilities,” the Elliptic report said.
Suggestions for protecting the crypto market from threats
The report says that the success of deterring the above threats depends on the degree of cooperation between various stakeholders. These include law enforcement investigators, cryptocurrency compliance officers, artificial intelligence and cryptocurrency users, technology developers, and policy makers and regulators.
The report recommends preventive measures, collectively referred to as DECODE.
Detect
Blockchain analytics should be used to identify the source of payments to illegal AI-related services – this can help track down offenders who use services such as unethical GPTs, generators of explicit AI forgeries, and services that provide documents to open fraudulent accounts on cryptocurrency exchanges and other services. Blockchain analytics enhanced by artificial intelligence will help identify instances of crime.
Educate
There is a need to raise awareness among cryptocurrency and AI users of both existing and recent signs of fraud, including the use of AI-related jargon to promote fraudulent investment platforms and fake AI-related crypto tokens.
Users and employees should be trained on how to identify “deepfakes” in potentially malicious posts on social media and video streaming platforms.
Co-operate
There should be data sharing to empower relevant stakeholders to combat AI-enhanced crypto crime.
Defend
The development team should consider that products and services that are resistant to potential criminal use from the outset can help prevent trends from spreading and avoid costly and stringent regulations in the future. Compliance teams should be formed, they will help to effectively detect, monitor and mitigate potential threats as these emerge by accessing blockchain analytics capabilities based on reliable data.
Enforce
Measures to combat illicit services experimenting with AI should be prioritized and ensure that new and rapidly evolving AI innovations are integrated with capacity building and training – this can enable effective detection and timely investigation of misuse of new technologies.
