Loading prices…
STKR NewsSTKR News0 of 3 free this month
Solana News

AI Will Make the Next 12 Months 'a Slaughterhouse' for Crypto Security: Anchorage Exec

The barrier to entry for high-level crypto hacks just hit zero. As AI scales social engineering, founders must rethink security before the next 12 months become a slaughterhouse.

Originally on Decrypt →
AB

Adrian Boysel

Contributor

Nov 18, 2026

4 min read

Photo illustration / STKR News

We have spent years worrying about smart contract bugs and reentrancy attacks. We built better auditors, launched bug bounties, and developed formal verification tools. But while we were hardening the code, we left the front door wide open. The consensus coming out of the recent Solana Breakpoint conference isn't about a new scaling solution or a better DEX aggregator. It is a warning: the next year is going to be a bloodbath for crypto security, and it is entirely thanks to AI.

The End of the Script Kiddie

For a long time, sophisticated state-sponsored attacks, like those attributed to North Korea's Lazarus Group, were identifiable because they were resource-intensive. If you were being targeted by a nation-state, you usually knew it because the social engineering was meticulous. If you were being targeted by a random scammer, you knew it because the grammar was terrible and the phishing link was obvious.

AI has deleted that distinction. Security experts are now sounding the alarm that Large Language Models (LLMs) have effectively given every low-level hacker a world-class social engineering department. You no longer need to speak the language or understand the cultural nuances of a target to manipulate them. You just need a prompt.

What we are seeing is the industrialization of the "human exploit." When we talk about a slaughterhouse in the next twelve months, we are talking about a massive spike in the volume and quality of phishing, deepfake voice cloning, and hyper-personalized outreach that makes the old-school Nigerian Prince emails look like cave paintings.

Why Builders are Unprepared

As founders, we tend to hire for technical competence first. We want the best Rust developers or the most efficient Solidity engineers. We assume that if the code is secure, the project is secure. This is a dangerous fallacy. Most modern crypto heists don't happen because someone found a zero-day in the protocol; they happen because a developer was tricked into downloading a malicious file during a fake job interview or because a CTO's voice was cloned on a Telegram call to authorize a move from a multisig.

The threat is no longer just about code. It is about the people who hold the keys. AI allows attackers to research your entire staff, understand their relationships, and mimic their communication styles in seconds. If an attacker can generate 10,000 perfectly written, context-aware phishing emails for the cost of a few API credits, the math of security changes forever.

The North Korea Factor

It is no secret that North Korean hackers have been treating the crypto industry like a personal ATM. What is changing is their efficiency. Security leads are reporting that these groups are now using AI to bypass the traditional red flags we teach employees to look for. They are creating convincing fake LinkedIn profiles, participating in developer communities for months, and using AI-generated avatars to conduct video interviews.

This isn't just a problem for big exchanges anymore. If you are a mid-sized DeFi protocol with $50 million in Total Value Locked (TVL), you are a high-priority target. The effort required to attack you has dropped significantly, while the potential payout remains massive.

Moving Beyond Multi-Sig

If the next year is going to be as bad as the experts suggest, the current industry standards for security are insufficient. A standard 2-of-3 multi-sig is a joke if the attacker can use AI to socially engineer two of the signers simultaneously. We have to move toward a culture of "zero trust" that actually means something.

  • Verify via Multiple Channels: If a team member asks for a sensitive move, verify it on a completely different platform using a pre-arranged code word that hasn't been typed into a computer.
  • Hardware or Bust: There is no excuse for not using hardware security modules (HSMs) or physical security keys for every single entry point. If it is software-based, AI can eventually find a way to spoof it.
  • Deepfake Awareness: Assume that any video or audio call could be faked. This sounds paranoid, but in a world where real-time voice cloning is a consumer product, paranoia is a feature, not a bug.

The Developer Perspective

From a builder's standpoint, this shifts the burden of security from the DevOps team to the HR and Culture teams. You have to train your people to be skeptics. This is hard because the ethos of Web3 is often about open collaboration and permissionless networking. We like to jump on calls with strangers to talk about partnerships. We like to hire pseudonymous developers from Twitter.

AI is weaponizing that openness. We are entering an era where the most dangerous thing you can do is trust a well-written DM or a familiar-sounding voice. The slaughterhouse metaphor is apt because, in a slaughterhouse, the victims don't see the end coming; they are led through a process that feels routine until it is too late.

The goal for the next twelve months isn't just to ship features; it is to survive the most sophisticated social engineering environment in the history of computing.

The Hard Truth

We are going to see some legendary names in this space get wiped out. Not because their math was wrong, but because they weren't prepared for the scale of AI-driven deception. If you aren't currently auditing your internal communication protocols and the way your team handles sensitive data, you are already behind.

The barrier to entry for hackers has disappeared. The tools are free, the targets are rich, and the defense is currently built on outdated assumptions. It is time to stop thinking about security as a line of code and start thinking about it as a psychological battlefield. If you can't protect your people from a machine that can talk exactly like you, you can't protect your protocol.


Read the original at Decrypt →

The Brief

Stay Updated on Cutting-Edge Tech

A six-minute morning dispatch on the markets and the technology shaping them.

Free. No spam. Unsubscribe anytime.

Write for STKR

Become a Contributor

Earn $STKR for published stories on markets, protocols, and culture.

  • Earn $STKR for every published piece
  • Editorial support from the STKR desk
  • Byline visibility across the network
  • First look at the upcoming creator program
Apply to Write

Keep reading

All stories

Comments

24 reader responses