A16Z backs blockchain tech to become critical to vetting AI agents and impersonation

AI agents may need blockchains to solve the issue of impersonation and verification. As agents have shown the ease of colonizing social media, verification is becoming more important. 

AI may be a key application area for blockchains, especially when it comes to user and identity verification, noted A16Z in a forward-looking analysis. While the Internet was made for human interaction, culminating in social media, AI agents have made it faster and easier to coordinate and produce content. 

The activity of AI agents may soon become indistinguishable from human activity, producing content at scale and linking across platforms. Human users are additionally required to solve captchas for verification at multiple steps. 

AI agents must be separated from human users

Separating AI agents from human users is not yet a part of the Internet toolset, nor does it create friction for humans. According to A16Z, blockchains can solve this issue by several approaches. The key ability is to create an immutable, tamper-proof identity that can be used across platforms. 

In the past, on-chain projects have attempted to create passport-like identities. For the first time, those identities will have a use case against AI agents on social media. 

AI can copy and create fakes of voice messages, faces, writing styles, video, images, and even an entire social media presence. One actor can also spin up multiple social media participants at a small extra cost. 

Decentralized proof of personhood can make it impossible to fake multiple personalities, not without proof of the original identity. Blockchains can limit the supply of generated IDs, thereby increasing the costs for attackers, or block them entirely due to the lack of proof of humanity. 

Chains can be used to prove human uniqueness, similar to Worldcoin’s technology. Identity may become the real scarcity, noted A16Z. 

Blockchains can give agents their own internet

Agents have shown they can, in theory, use the Internet on their own. Projects like Moltbook show the potential usage for agents, especially in building communication similar to social media. As Cryptopolitan reported, the inclusion of AI agents requires additional risk management to prevent exploits and scams.

According to A16Z, agents may also have their unique identities to prove themselves across platforms. An on-chain identity layer, possibly based on tokens or NFTs, can give agents a unique universal passport. The identities can also carry data on capabilities, permissions, and payments, and can be verified from multiple points on the Internet. 

The identity will prevent agent spoofing, allowing for the creation of more useful AI assistants. 

Additionally, blockchains are already built as an almost ideal environment for AI agents. As more attention is paid to agentic transactions and payments on behalf of humans, the on-chain environment becomes more meaningful. 

Existing on-chain tools already allow for AI-native payments, including near-zero cost transactions, high-velocity payments, micropayments, and smart contracts. Blockchains can turn into a machine-native venue, capturing value that is not within the reach of humans.

AI can add another layer to blockchain automation, going beyond simple bots for specific tasks.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.