Urbelis expressed his belief that artificial intelligence could fundamentally transform the standard practices in smart contract development. Traditionally, development teams have cited the high costs and intricate nature of audits as reasons for not conducting certain evaluations. However, this rationale becomes less viable when advanced security analysis tools are readily accessible.
“An AI-generated report won’t hold up as a protective measure,” he commented. “A claimant might argue the opposite: since the tool was available and affordable, it should have identified the issue.”
This situation sparks larger discussions within the sector: if AI-driven security assessments become commonplace, will investors demand these evaluations before financing projects? Furthermore, could the failure to conduct AI-assisted audits be perceived as a form of negligence?
Even with the potential advantages of this technology, both experts maintain that AI is not likely to completely supplant human auditors.
While machines are adept at detecting coding errors, Urbelis pointed out that they tend to struggle with recognizing economic and incentive-related vulnerabilities that have led to major financial losses in the crypto space. “The issues that drain funds often involve intent and adversarial motivations,” he explained. “Those intricacies still require a skilled human presence.”
Schwed echoed this caution. “‘Claude, please audit my smart contract without errors’ does not constitute a robust security strategy,” he noted. “If the operator of the tool cannot interpret the results, you haven’t secured safety; you’ve merely acquired a misleading sense of it.”