Has artificial intelligence (AI) killed the metaverse? Large Language Model (LLM) AIs like ChatGPT have sucked up all the oxygen (and all the VC investment) in the technosphere since November 2022. ChatGPT hit 100 million monthly active users in January of 2023, less than two months after launch. By contrast, Meta’s virtual reality (VR) platform, Horizon Worlds, finished 2022 with under two hundred thousand active users—less than it started the year with. Still, anyone writing a metaverse obituary should cap their pen: for AI companies, many of which have yet to secure revenue streams to pay for intense cloud resources their applications require, the underpopulated metaverse is just unclaimed territory. Get ready for the land rush.

Pretty soon, you’ll be able to transport an object to the metaverse as a 3D asset by taking its picture or create an object by describing it. Not just objects, either, but characters. People. Already, experiments by Stanford and Google have shown that ChatGPT can create autonomous characters with distinct personality traits that will behave in shocking human-like ways. They can pursue goals, socialize, and gossip. If Meta is concerned its VR community feels like a ghost town, they now have the means to populate it.

These AI metaverse dwellers will have a lot of beneficial applications: more immersive and engaging game NPCs, tutors and trainers, personal assistants, etc. The potential for abuse of this technology is also so obvious as to be inevitable. Social scams, which rely on cozying up to a target over various social platforms in order to eventually get them to give the scammer access to their crypto wallet or other financial data, are already a multi-million-dollar industry. Imagine how fast this sector of the scam economy will grow once it can be fully automated.

LLMs are already way past the Turing test; it’s unreasonable and unfair to expect individuals to be able to defend themselves from this kind of attack. The metaverse is going to need systemic solutions to prevent the rise of an infinite army of inexhaustible, automated scam artists. We can look to blockchain to understand both the risks and the solutions: in crypto, transactions are technologically very secure, but especially susceptible to social engineering scams. If a bad actor can convince a person to sign a malicious smart contract, they can drain all their assets instantly—or, plant a backdoor that lets them quietly siphon away tokens over time. The only way to protect crypto users is a zero-trust approach, inspecting the origins and history of every transaction, contract, and address for signs of fraud. Users can enable these kinds of protections themselves in their own crypto wallets, but the more effective and farther-reaching approach is for crypto projects to build these background check and authentication tools into their platforms by design.

AI, far from killing the metaverse, is going to help transform it from an empty landscape to a fully realized digital universe. In the process, we’re going to travel through some wild and unclaimed territory. When you don’t ever know whether you’re talking to a real person or an automated scambot, the phrase “trust but verify” takes on a whole new meaning. The metaverse needs AI to grow—but it’s going to need powerful security tools and built-in user protections to contain the potential of AI to wreak havoc.

Editor’s Note: This blog is part of a series for Artificial Intelligence (AI) Appreciation Day, which is held annually on July 16. Click here to read more AI Day stories from LAVNCH [CODE] and click here to read more AI Day stories from rAVe [PUBS].

About the Author: Brittany Mier y Terán is head of business development for Harpie, the company that created the first on-chain firewall preventing hacks, scams, and theft. She specializes in enterprise grade blockchain security solutions, and is passionate about public goods projects that focus on onboarding the next generation to Web3. Previously a leader in the consumer technology industry, she was recently listed as one of the 40 Under 40 and honored with the title “Woman to Watch” by Women in Consumer Technology at CES 2022. Mier y Terán is an avid ETHGlobal hackathon participant, competing alongside her two daughters, where they have won multiple bounties from Polygon, Lens, WalletConnect, Optimism, Chainlink, and LivePeer.