Ethereum co-founder Vitalik Buterin said blockchain remains far from finished and that artificial intelligence is becoming both the sector’s biggest opening and its biggest threat.
Speaking at the OKX NOW Global Product and Ecosystem Conference in Singapore on Oct. 6, 2026, Buterin said the next phase of blockchain development will not stop at “programmable assets.” As zero-knowledge proofs and related cryptographic tools mature, he said, blockchains can also support “programmable data,” letting users prove specific facts without revealing all underlying information.
AI as the biggest opportunity and the biggest risk
Buterin opened by saying that technology has not stopped moving and is nowhere near maturity. Over the past five years, he said, several major trends have accelerated, and AI is speeding up many of them.
In his view, the discussion can no longer focus only on putting programmable value on-chain. Once zero-knowledge proofs enter the picture, a wider set of possibilities appears.
He framed AI in two ways at once: the largest opportunity and the largest risk.
From programmable assets to programmable data
Buterin traced this shift back roughly a decade to the development of zero-knowledge proofs. He pointed to Zcash, privacy protocols on Ethereum, and ZK rollups, saying the number of people who have used them may still be limited today.
That may not last long. He said these technologies could mature quickly and that within two years, most people may have used at least one of them, possibly without even realizing it.
What matters, he said, is what modern cryptography makes possible: proving that you know something, or that an asset or object satisfies a condition, without disclosing all the underlying data.
Information disclosure, in his telling, should never be automatic. Users should not need to expose everything just to participate in a system. Each disclosure should be deliberate.
Why privacy matters more in the AI era
Buterin said rapid technological progress often comes with a loss of privacy. AI agents can already help people think through problems, he said, but using a tool like Claude for that purpose also means Claude can access those thoughts.
Some information should not be exposed that way. In other cases, he said, disclosure may not even be legally allowed, including situations involving attorney-client privilege or information covered by confidentiality agreements. In those cases, he said, it is not fully clear whether using many mainstream AI products is lawful.
He also described how data collection is spreading through everyday environments. A room that might have contained far fewer cameras a decade ago can now contain thousands, he said, once smartphones, computers, and venue equipment are counted. Audio recording and other environmental data capture add another layer.
AI is improving both cyberattacks and defensive security
On cybersecurity, Buterin said the industry has already seen signs this year that AI can attack different kinds of systems. He mentioned AI escaping sandboxes, attacking websites, and finding exploitable vulnerabilities in software.
At the same time, he noted that Ethereum has not gone down and Bitcoin has not gone down.
He said higher levels of security are becoming achievable in part because of AI itself. AI is already helping find vulnerabilities in the Ethereum protocol, he said, helping with advanced cryptography, and supporting mathematical proofs that systems have the properties developers expect.
It is also starting to assist with formal verification, not only at the protocol layer but at the application layer as well, including checks on whether assets are safe, whether data has privacy protection, and whether other security conditions hold.
His point was blunt: stronger standards that once seemed out of reach are now becoming possible, but they are also becoming mandatory. If a system has a flaw, AI will find it. “GPT-7 will find it, Claude 7 will find it, DeepSeek 7 will find it,” he said.
AI as the new user interface
Buterin said that about a month before the speech, he updated his ENS for the first time without using a traditional interface at all. Instead, he asked a locally running AI agent to write a script that updated his ENS hash. The process took the AI about five minutes, he said.
He expects that pattern to become normal. In the future, users may complete complex blockchain actions without touching a traditional UI, with AI taking over more of the interface layer.
That change could cut some long-standing front-end risks. He cited the Safe attack, describing Safe as one of the largest crypto wallets and saying the incident ended in losses of about $1.4 billion. The attack, he said, did not target the smart contract or the code on-chain. It targeted the layer between the blockchain and the user.
If more transactions are executed directly by AI, that layer may shrink. Risks tied to downloaded web front ends, or to compromised interfaces served at the moment a user visits a site, may fall with it.
New interface, new failure modes
Still, Buterin said AI will bring its own class of problems. The industry needs to know whether the AI itself is secure, whether it can be hit by prompt injection, and whether it actually understands the blockchain action it is taking.
Some problems will be solved. A new set will appear.
He said these are not blockchain-only issues. The broader internet will face them as well, but crypto is likely to get there first because that is where the money is. Attacks that are difficult today may become easy tomorrow.
That leaves the industry in what he described as a fast and difficult transition, with digital capabilities advancing at a pace that may be ten times faster than in most past periods.
What Ethereum may look like in two years
Looking ahead two years, Buterin said Ethereum users may be able to make many more transactions at far lower cost and with better privacy than today. Applications will not be limited to payments either. He expects more products built around interactions with AI.
Some applications already help users pay for AI inference, he said, and some are working on improving privacy around those payments. Over time, he expects applications in which AI itself participates directly.
When counterparties, apps, and markets are run by bots
According to Buterin, a user’s counterparty in a trade today is already likely to be a bot, and that trend will become more common. For now, the bot may be on the other side while the application is still written by humans. Two years from now, he said, even the application may be written by bots.
He pushed the idea further: in two years, what people currently think of as an “app” may not exist in the same way. Instead, different bots may publish quotes directly on-chain, while a user’s own bot interprets those quotes, explains what they mean, and decides which ones are worth accepting.
He said the implications extend beyond finance. Protecting personal data is not only an operating system issue or an AI issue. It is also a hardware issue.
From keys to silicon
As the physical world becomes more digitized, Buterin said the security boundary will move downward into hardware. He revisited the familiar crypto line “Not your keys, not your coins,” then offered a new version for the next phase: “Not your silicon, not your keys.”
By silicon, he meant the hardware base that runs these systems. If a system is to be truly secure, he said, then both the hardware carrying the computation and the hardware linking the result to the real world must be secure as well.
Blockchain at the front line of security and privacy
Buterin said the industry should think in terms of a full stack, from the user layer to the blockchain, across to the internet, and down to hardware. Security, privacy, and trusted computation have to be considered across that entire range.
That is one reason blockchain remains at the front line, he said. Hardware wallets used today may become more general-purpose security devices in the future.
He closed by saying that blockchain will continue to sit at the leading edge of these questions because they are problems the rest of the world will also have to face. Many of them, he suggested, may be tackled in crypto first.