{"id":91236,"date":"2025-11-20T14:51:25","date_gmt":"2025-11-20T11:51:25","guid":{"rendered":"https:\/\/forklog.com\/en\/?p=91236"},"modified":"2025-11-20T14:56:22","modified_gmt":"2025-11-20T11:56:22","slug":"zkps-are-the-foundation-of-a-new-ai-era-says-billions-networks-ceo","status":"publish","type":"post","link":"https:\/\/forklog.com\/en\/zkps-are-the-foundation-of-a-new-ai-era-says-billions-networks-ceo\/","title":{"rendered":"ZKPs are the foundation of a new AI era, says Billions Network\u2019s CEO"},"content":{"rendered":"

The technology of zero-knowledge proofs (ZKPs)<\/a> will underpin a new era of trustworthy artificial intelligence and digital identity, says<\/a> Billions Network CEO Evin McMullen.<\/p>\n

It will allow individuals and organisations to interact with platforms securely and transparently.<\/p>\n

The problem<\/h2>\n

As AI agents proliferate, people are asked to hand over ever more biometric data to prove they are not bots.<\/p>\n

If personal information leaks into the open, machines can exploit it\u2014for instance to impersonate real people. That can undermine systems designed to filter out \u201cnon-human\u201d users.<\/p>\n

An arms race ensues: the more rigorous the checks, the higher the risk of data spilling onto the internet.<\/p>\n

It is unacceptable to demand transparency from people while tolerating opacity from machines, McMullen argues. Both bots and online users need more effective ways to verify identity.<\/p>\n

Hoovering up biometrics and building centralised registries is no solution, since such databases inevitably attract cybercriminals.<\/p>\n

A trust deficit stymies progress<\/h3>\n

The absence of mechanisms to reliably verify the identity of artificial intelligence creates problems. So long as AI agents can pass as humans, manipulate markets or execute unauthorised transactions, enterprises will be slow to roll out autonomous systems at scale.<\/p>\n

McMullen stressed that LLM<\/span>s become more hazardous after fine-tuning\u2014they are 22 times more likely to produce harmful or undesirable responses than the original \u201cbase\u201d models.<\/p>\n

Such fine-tuned systems are also easier to jailbreak, that is, to circumvent built-in safeguards.<\/p>\n

If a service cannot tell who is calling it, any interaction with AI becomes a potential data-leak or misuse risk.<\/p>\n

Legitimate AI systems need verifiable credentials to participate in the emerging \u201cagent-to-agent\u201d economy. If a trading bot transacts with another bot, both sides must be confident in the counterparty\u2019s identity and system of accountability.<\/p>\n

For people, the status quo is also flawed. Traditional identity checks often lead to mass data breaches, enable authoritarian surveillance and yield billions in corporate revenues from selling personal data with no compensation to the owners.<\/p>\n

People do not want to share personal data, yet regulatory obligations force them to.<\/p>\n

The solution<\/h2>\n

Zero-knowledge proofs could become the tool that lets both humans and AIs attest to their privileges without inviting abuse.<\/p>\n

ZKPs let parties validate certain claims without revealing the underlying data. Examples:<\/p>\n