Firm using cryptography to keep AI accountable joins World via TFH acquisition

Tools for Humanity (TFH), which calls itself a “contributor” to the World ID iris biometrics and digital identity project, has a new limb. A Substack post from Modulus Labs’ Daniel Shorr says his company is joining TFH “to build the largest network of real humans on the planet.”
Modulus Labs reportedly developed its product, Accountable Magic, in response to the increasing dysfunctionality of the internet: “misinformation, bots and scammers, the anger economy.” Their cryptographic system is “a way to prevent the manipulation of AI algorithms — mathematically.” Specifically, it focuses on “machine-learning accountability” using zero-knowledge proof (ZKP) protocols.
Now, it will be integrated into World’s system for collecting iris biometrics and creating a World ID, thereby recording and verifying “humanness.”
‘Accountable AI’ system like a blue check for algorithmic content
According to its website, Accountable Magic provides “AI security through novel cryptography.” In practical terms, it is an edge-based system to verify that AI algorithms have not been manipulated – “like Twitter’s blue check but for AI outputs.”
Per Shorr’s blog, “zero-knowledge AI will deliver private and verifiably secure authentication on the user’s personal device, playing an important role through Personal Custody.”
“Along with the Orb and innovations like AMPC (anonymized multi-party computation), this technology will play an important role in distinguishing between bots and humans at the scale of billions. That means online interactions that are safer and more accountable. It means digital authorship that’s verifiable. And it means bringing new communities to the astonishing power of digital currencies.”
Use AI to fight AI to enable further use of AI…
A 2023 Fortune Crypto article about the startup says that through zero-knowledge proofs, “outside observers can verify that companies or developers used a promised AI algorithm. For example, OpenAI, the juggernaut that developed ChatGPT, can prove that its chatbot wrote a poem without revealing the algorithm’s ‘weights,’ or what an A.I. model learns after training on copious amounts of data.”
Its proposition, then, is that using cryptography to keep algorithms accountable will enable developers to “build wildly expressive services that never betray our trust.” A graphic on its site cycles through use cases: NFT appraisals, private identity authentication, AI game economies, “tamper-proof intelligent finance,” “authentic machine artists.”
And so do forces continue to muster in the escalating war between AI algorithms that could lead to harm and tech tools to temper them – so that even more exciting algorithms can be developed.
Article Topics
Accountable Magic | acquisitions | digital ID | digital identity | Modulus Labs | Tools for Humanity | World | zero knowledge
Comments