MITNews
Q: Why do we’d like proof of identity?
Tobin South: AI capabilities are improving rapidly. While much of the general public discussion is about chatbots improving, advanced AI enables much more capabilities than simply higher ChatGPT, comparable to the flexibility for AI to interact autonomously online. AI could have the opportunity to create accounts, publish content, generate fake content, impersonate a human online, or algorithmically amplify content at scale. This poses many risks. You can consider this as a “digital imposter” problem, where it becomes increasingly difficult to differentiate between advanced AI and humans. Identity verification is one potential solution to this problem.
Nour Soliman: Such advanced AI capabilities could help criminals perform large-scale attacks or spread misinformation. The web might be stuffed with AIs that regurgitate content from real people to run disinformation campaigns. It will turn out to be harder to navigate the web, and social media specifically. One could imagine using identity verification to filter out certain content and moderate content in your social media feed or determine the trustworthiness of knowledge you receive online.
Q: What is proof of identity and how are you going to make sure the security of such proof?
South: With a proof of person, you may prove that you just are human without revealing the rest about your identity. With this proof, you may get information from an entity like the federal government that may guarantee that you just are human, after which use privacy technology to prove that fact without revealing sensitive details about your identity. To get a proof of person, you have got to indicate up in person or have a relationship with the federal government, like a tax ID number. There is an offline component. You must do something that only humans can do. AIs can't show up on the DMV, for instance. And even essentially the most sophisticated AIs can't forge or break cryptography. So we're combining two ideas – the safety we now have from cryptography and the undeniable fact that humans still have some capabilities that AIs don't have – to create really robust guarantees that you just are human.
Soliman: But identity credentials may be optional. Service providers can leave it as much as users whether or not they wish to use one or not. If people currently only wish to interact with real, verified people online, there's no reasonable option to try this. And beyond just creating content and talking to people, AI agents will eventually act on behalf of individuals too. If I would like to purchase something online or negotiate a deal, I’d wish to make sure I'm interacting with entities which have identity credentials to make sure they're trustworthy.
South: Identity verification builds on an infrastructure and set of security technologies which have been available to us for many years, comparable to using identifiers comparable to an email account to log into online services, and it could possibly complement these existing methods.
Q: What risks are related to identity verification and how are you going to reduce these risks?
Soliman: One risk comes from the best way identity proofs is likely to be implemented. There are concerns about concentration of power. Let's say a certain entity is the only issuer, or the system is designed to present all power to at least one entity. This could cause lots of concern amongst a portion of the population – perhaps they don't trust that entity and don't feel secure working with it. We have to implement identity proofs in a way that folks trust the issuers and be sure that people's identities remain completely isolated from their identity proofs to guard privacy.
South: If the one option to get proof of identity is to physically go somewhere and prove that you just are a human, then that might be scary in case you are in a sociopolitical environment where it’s difficult or dangerous to go to that physical place. That could prevent some people from freely sharing their messages online, potentially limiting free expression. That is why it is vital that there’s quite a lot of issuers of proof of identity and an open protocol to be sure that freedom of expression is maintained.
Soliman: Our paper goals to encourage governments, policymakers, leaders and researchers to take a position more resources in personhood credentials. We suggest that researchers explore different implementation directions and examine the broader impact that personhood credentials could have on the community. We have to be sure that we create the suitable policies and rules for the implementation of personhood credentials.
South: AI is evolving in a short time, actually much faster than the speed at which governments can adapt. It's time for governments and enormous firms to begin enthusiastic about how they’ll adapt their digital systems to have the opportunity to prove that somebody is human, in a way that’s privacy-preserving and secure, in order that once we reach a future where AI has these advanced capabilities, we’re ready.
