Blockchain analytics must improve to hold up in court

Blockchain analytics providers must standardise data and improve their tools to withstand scrutiny in court and assist cross-border investigations, according to global law enforcement leaders. Providers should standardise data taxonomies and attribution to be interoperable with other systems, the Blockchain Intelligence Forum heard this month at the Digital Innovation Summit in Bucharest.

“Where do you make that transition from blockchain intelligence to actual evidence that will stand up in court? For this we need to take a step back and look at what we are actually doing when we’re analysing blockchain transactions and blockchain data,” said Daniel Leon, cryptocurrency specialist at Europol’s EC3 (digital support unit and cyber intelligence).

Transactions recorded on the blockchain can be considered almost forensic in nature and are almost never questioned in court. “It’s the interpretation of that data when things start to be called into question,” Leon said.

Every time blockchain analytics firms get a result, they should be asking: “how do we know it’s right?”, said Patrick Tan, general counsel at ChainArgos in Singapore.

“We must highlight the importance of developing robust blockchain intelligence methodologies that are rigorously tested and verified. There is no regulation in this industry,” Tan told the forum.

“My opinion is just as valid as yours; that doesnʼt necessarily make mine the truth. We have a rare and unique opportunity to get to that truth, but we need to embrace blockchain technology as a data science task, not one thatʼs made up of hope and opinions,” he said.

Overwhelmed by data

Police are overwhelmed by the amount of data to be processed and analysed, and in turn investigators and analysts require a huge amount of expertise and technical knowledge to sift through the data using artificial intelligence (AI) and blockchain analytics tools that use AI. These tools bring advantages, but also present their own challenges.

A major problem facing law enforcement is interoperability of data, said Vincent Danjean, head of Interpol’s cyberspace and new technologies laboratory in Lyon, France.  

“We use the solution providers who are dumping blockchain data into AI, or scrape darknet content to dump it into AI and large language models. Investigators now have to rely on those AI-aided analyses to make sense of their cases. With AI-aided pattern analysis, we may now classify certain activities on the blockchain as suspicious or related to money laundering, for instance,” Danjean said.

One problem law enforcement is facing with AI is the a lack of explainability in the processes behind the results its produces or the ‘black boxʼ effect.

“We cannot reproduce and explain what AI is doing under the hood. In a judicial procedure, that typically means you would fail your case because the defence would use a different solution and come up with different answers,” Danjean added.

He said for example, if law enforcement in one country is investigating a money laundering case and want to share data related to their findings with an agency in another country, it will be difficult to analyse and compare if they use blockchain analytics providers: connections made between the wallets, the clusters, the transactions and the findings from one country will be different. Wallets tagged and attributed in some taxonomies as a money mule case, for instance, could be designated as drop wallets in another taxonomy.

“So how do you make connections between the two?” Danjean said. “How do you make sure you are exhaustive in your investigation?”

Tag packs

Some blockchain intelligence companies are already working with law enforcement to standardise data. In collaboration with Interpol, Austrian blockchain intelligence company Iknaio Cryptoasset Analytics has devised a standardised data format called tag packs. This is an open standard, which is a file format that sets out an agreed minimal set of fields and also specifies how the values should look in those fields. Iknaio has also developed a standardised taxonomy for service types and corporate entities as well as different types of abuses and typologies.

“Why are we doing this? Because we want evidence basically provided in front of court to be court-proof, and you have to build some basic requirements like reliability and verifiability — which I have to say is currently often not [available] given the current blockchain intelligence space,” said Bernhard Haslhofer, Iknaio co-founder and crypto finance research group lead at the Complexity Science Hub in Vienna.