The Facts
On April 19, 2026, crypto investor and founder of Bankless, Ryan Sean Adams, publicly warned that artificial intelligence is rapidly changing the threat landscape for digital assets. His comments followed a major exploit involving Kelp DAO, where attackers drained approximately $292 million from the platform, making it one of the largest decentralized finance breaches of the year.
The attack itself is still under investigation, with no confirmed evidence that AI tools were directly used in executing the exploit. What is clear is that the breach relied on familiar weaknesses within decentralized finance systems, including smart contract vulnerabilities and complex protocol interactions that can be difficult to fully secure at scale.
Adams’ warning did not focus on the mechanics of this specific hack alone. Instead, he pointed to a broader shift, arguing that AI is giving malicious actors new capabilities to identify vulnerabilities faster, automate attacks, and operate at a level of sophistication that was previously harder to achieve.
The Blame
The framing came quickly and clearly.
In his public statements, Ryan Sean Adams described AI as giving hackers what he called “dark superpowers.” It is the kind of phrase that travels well, and it did. Within hours, the narrative began to take shape: crypto is under attack, and AI is accelerating the threat.
It is not an entirely baseless concern. AI tools can assist with code analysis, pattern recognition, and automation. But the leap from capability to causation is where things become less precise. The hack happened. The losses are real. The connection to AI, at least in this case, remains speculative.
Still, the appeal of the explanation is obvious. It shifts the conversation away from platform design, security practices, and risk management, and toward something larger, more abstract, and harder to pin down.
The Real Story
Strip away the framing, and the structure looks familiar.
A decentralized finance platform lost hundreds of millions of dollars due to vulnerabilities that, while complex, fall within a known category of risks in the crypto space. These systems are often built for speed and innovation, sometimes at the expense of exhaustive security testing. When they fail, the consequences are immediate and irreversible.
There is no confirmed evidence that AI wrote the exploit, executed the attack, or made the decision to target Kelp DAO. What exists instead is a system with identifiable weaknesses and actors who took advantage of them.
Attributing the outcome to AI may capture the mood of the moment, but it does not explain the mechanics of what happened. The responsibility still traces back to human design choices, human oversight, and human exploitation of those gaps.
The Aftermath
In the days following the exploit, discussions across the crypto industry have shifted toward security, audits, and the growing complexity of defending decentralized systems. Platforms are under renewed pressure to strengthen protections, while users are once again reminded of the risks tied to high-yield protocols.
At the same time, the narrative around AI continues to gain traction. As tools become more accessible and more powerful, they are increasingly cited as a force multiplier in both innovation and exploitation.
Whether or not AI played a role in this specific incident, it is already becoming part of the explanation for future ones.
The Verdict
WHO’S BLAMING AI:
Ryan Sean Adams, who warned that AI is giving hackers “dark superpowers” following the Kelp DAO exploit.
WHAT ACTUALLY HAPPENED:
A decentralized finance platform suffered a $292 million breach linked to known vulnerabilities. There is no confirmed evidence that AI directly enabled the attack.
WHO GOT AWAY WITH IT:
The attackers remain unidentified, while broader responsibility shifts toward an emerging narrative about AI rather than the specific weaknesses that were exploited.
BLAME RATING: 🤖🤖🤖 (3/5 robots) – AI may be part of the future threat landscape, but in this case, the evidence points to familiar human-made vulnerabilities doing what they have always done.
