On Valentine’s Day, Indonesia’s citizens participated in an unprecedented global democratic wave . More than four billion people could vote across 60 countries in 2024, marking a record year for general elections worldwide.
This heightened global electoral activity raised concerns about the integrity of democratic processes amid emerging threats, including that posed by artificial intelligence (AI).
The shadow of past events, such as the Cambridge Analytica scandal, which saw private data from 87 million Facebook users harvested without consent , loomed large. Since then, the introduction of technologies like Apple’s Vision Pro and platforms such as ChatGPT led to an unparalleled level of personal data collection and potential misuse.
In light of these developments, CCN interviewed Harry Halpin, the co-founder and CEO of Nym Technologies. Halpin shared insights on how AI could influence genuine democratic engagement. He also offered guidance on safeguarding against the potential manipulation of democratic values.
Halpin: Currently, digital privacy is a relatively unexplored and underfunded area. The reason is twofold. First, the current Web 2.0 is based on a surveillance-based business model , which includes both advertising and selling user data. Second, it is naturally harder to develop privacy-enhanced and secure technologies, which is one reason they were not part of the original Web and Internet. This means the Internet and Web cannot be easily used for democratic purposes, including e-voting, but is instead used for behavior prediction and control.
Halpin: I was surprised people were even surprised by Cambridge Analytica . It was common knowledge that Facebook was abusing its user data at the time. I do not consider people like Brittany Kaiser to be whistleblowers, unlike Edward Snowden, who revealed top secret information and US government mass surveillance by the NSA. Likewise, it should be completely unsurprising that Apple VisionPro is tracking users – even their eyeballs – and ChatGPT is recording and selling the conversations their users are having, which many users would consider private.
Halpin: Data mining – another name for machine learning or AI – will be used to target advertising and to modify voting behavior. This is used in both the US and Europe by all parties. For example, it should come as no surprise that David Recordon , who invented the Like Button, worked for the Obama Administration after leaving Facebook.
So we should expect subtle changes in algorithmic feeds on platforms like Facebook in order to influence voting behavior . This is dangerous because it’s hard to detect, even by experts. The entire purpose of algorithmic data-driven targeted advertisements – which are often not declared as such – is to subconsciously influence behavior, which is just considered mind control.
Halpin: There has always been propaganda, but we can expect increased polarization due to behavior manipulation. I would say all sides will launch misinformation, both Biden and Trump. A quick look at the situation around Israel and Gaza shows the tremendous problems around detecting misinformation. The problem is that one person’s misinformation is another person’s objective news. No blockchain technologies we have seen are mature enough to help right now.
Halpin: The first step is using a VPN, such as our upcoming NymVPN or Tor. The second is using open-source software like Linux rather than Windows and Max, and keeping the software updated. The third is to use two-factor authentication. Then I would recommend using software like Signal for messaging and avoiding centralized crypto exchanges.
Halpin: Governments should ban surveillance and targeted advertisements. Surveillance and data-driven advertising are just as dangerous as nuclear weapons. Governments should mandate privacy-enhancing technologies.
Halpin: All technology should be privacy-enhanced and respect basic rights, just as cars have seat belts nowadays. While almost all blockchain-based voting schemes are scams or failures, there is well-explored academic literature in using mixnets for electronic voting. Blockchain-based integrity checking of data provenance can distinguish generated AI from content generated by humans.