Elliptic integrates ChatGPT to bolster crypto risk detection

Cryptocurrency risk management firm Elliptic has integrated ChatGPT to ramp up its efficiency in detecting crypto threats – a move which comes amid other crypto firms reporting mixed results in its implementation.

Elliptic provides risk assessments to crypto users regarding transactions, wallets and exchanges. These are conducted by its team of human researchers utilizing a proprietary dataset with “over a decades worth of data.”

In an effort to improve the accuracy and speed of identifying new risks, Elliptic has implemented ChatGPT into its off-chain intelligence and research operations, according to a June 1 statement.

It was noted that ChatGPT will allow its researchers and investigators to tackle new risk factors “in higher volumes and at a quicker speed than ever before.”

Jackson Hull, the CTO of Elliptic, suggested that its implementation has the potential to detect risks that human researchers may have failed to recognize:

“Our customers come to us to know exactly their risk exposure. Integrating ChatGPT allows us to scale up our intelligence, giving our customers a view on risk they can’t get anywhere else.”

Several crypto firms have implemented ChatGPT into its operations in recent times, yielding mixed results.

Cryptocurrency exchange Crypto.com launched a ChatGPT-based AI user assistant, Amy, on May 3, in an effort to help inform users about the industry, including real-time token prices, projects, and historical events.

While a Crypto.com spokesperson told Cointelegraph that AI has the potential for “massive potential implications” in the crypto industry, they expect a lot of feedback in the early days, which will be “integrated into future upgrades.”

However, for cryptocurrency derivatives exchange, Bitget, “cracks” have already been identified in the accuracy of ChatGPT’s responses.

Gracy Chen, managing director of Bitget, reported to Cointelegraph on June 1 that while Bitget was initially impressed by ChatGPT to handle routine customer inquiries, errors were spotted when given “more complex queries.”

“Despite being trained on extensive data, ChatGPT sometimes offers misleading, biased or incomplete information, as a recent internal test reminded us” Chen noted.

Related: OpenAI commits $1M to support AI-driven cybersecurity initiatives

These developments come after Solana Labs announced on April 26 that Solana (SOL) users will soon be able to interact with the Solana network directly from ChatGPT.

Through an open-source plugin created for ChatGPT, users will be able to check wallet balances, transfer tokens and purchase NFTs.

Solana Labs is yet to announce when this feature will be publicly available.

Cointelegraph reached out to Elliptic comment but did not receive a response by the time of publication.

Magazine: Musk’s alleged price manipulation, the Satoshi AI chatbot and more: Hodler’s Digest, May 28 – June 3

Article Categories: