The Ukrainian government announced on May 1 that they would use a “digital person” to represent the ministry of foreign affairs for at least some of their announcements. It (she?) will be a representative for the ministry on consular affairs and is called “Victoria Shi”, which refers to Ukraine’s goal of victory in their war with Russia (or perhaps rather Russia’s war on them) and the fact that it’s an artificial intelligence.

In the first message generated by/for this new representative, it represents itself as a “digital person”, and also “not a real person”.

The advantages for the ministry of foreign affairs to have a digital representative for consular affairs are clear enough: regardless of the war being waged on Ukraine, and regardless of any military assistance Ukraine may or may not receive, as long as someone pays for the computing power and software needed to run this entity, this “digital person” cannot be killed. It is immune to missile strikes, assassination attempts, etc. As long as the data center it runs out of remains standing, it will keep being. That data center is likely to be in a NATO country somewhere, so the likelihood of it falling victim to a physical attack from Russia is slim.

This “digital person” is, of course, vulnerable in ways that a human person would not be: a human person, even a “digitally enhanced” one, or even the digital “piece” or “avatar” of a human – what some have called a Homo numericus1, can only be truly destroyed by the physical death of that human person. While the digital identity of a person can be attacked, mutilated, and perverted through cyber-attacks in ways that weren’t possible before the advent of the notion of Homo numericus in the early 21st century, the “digital person”, more so than the digital identity of a human or legal person, can be utterly destroyed through cyber-attacks.

The innovation the Ukrainian ministry of foreign affairs announced a few days ago may go beyond the headlines that will doubtlessly last for only a few days, and the work of this particular “digital person” that will doubtlessly be retired in a few years, when Ukraine will hopefully have survived yet another attack on its existence. We’ve seen chat bots come and go before, but this one may be given a role that is beyond what any previous chat bots have been allowed to do: represent a ministry for consular affairs while real diplomats’ time is freed up to work on more important things. It is not clear to me to what extent Victoria Shi’s comments and statements will be checked and vetted before they are put out to the world, or how interactive this digital person will be, but the QR code at the bottom right of its videos will, at least to some extent, serve as the Ukrainian government’s stamp of approval. All software has bugs, and this will be no different, so at some point, this representative will say or do something to embarrass the Ukrainian foreign ministry. When that happens, though, the fall-out is unlikely to be any more severe than a junior diplomat putting their foot in their mouth. Young diplomats learn from that kind of experience. I don’t know of Victoria Shi will. I also don’t know if it matters if it is later replaced by a better-trained, more competent AI.

However, AI has also been known to hallucinate. This may make it a tempting target for cyber-attacks: an attack on the AI may get it to say things that, as confidence of the overseeing human grows and oversight lessens, may go unnoticed. I can easily imagine an AI like this making up facts and be found to be “lying”. Similarly, I wonder if the Ukrainian government will allow Victoria Shi to be interviewed by a journalist and, if so, if that journalist will dig into a topic it isn’t as thoroughly trained for 2.

If you were a Russian aiming to destroy the Ukrainian government by any means necessary, what would you do with this knowledge? How would you subvert Victoria Shi, and how would you go about destroying this particular little glimmer of hope? Obviously, cloud computing being what it is, you can’t attack it overtly beyond mocking it with your troll farms and spreading misinformation about it. You can, of course, use the likelness and voice samples of the same Ukrainian influencer to create a facimile of Victoria Shi and have its QR code point to a website that simply looks like the Ukrainian government’s website. That type of misinformation is something we know Russia already excells at, and has allowed them to influence politics, and I dare say policy, in Western countries in the past3. Beyond that, more aggressive cyber attacks are imaginable, where the underlying software or the data it pulls from is corrupted in ways to have Victoria be less irritating to its masters’ enemies.

But let’s put the war aside for a moment, and dip our toes into science fiction. AIs out-smarting the Turing test are a thing of the past, present, and future. While I don’t think the singularity is upon us, the thought experiment of a Homo numeris – a digital man, a man not belonging to, but emergent from, numbers, is an interesting one. Setting aside fanciful “quantum human computers”4 and the idea that human brains are, in fact, quantum computers5 serious work has been put into the idea that quantum computers could simulate consciousness6 7 8. What if they succeed? How would a conscious computer interact with the world? What rights and privileges would a digital person, a Homo numeris, have, if any? Are we as a species ready to enslave a species that we have created? Is Data in our future? If Data is in our future, is The Measure of a Man as well? Are we ready to open this particular Pandora’s box?

To be clear, I think that box is firmly shut for the moment: all the cloud computing in the world can’t create more than a facimile of an actual person, digital or otherwise. But the Ukrainian government has given us a glimpse of what a future with actual digital people might look like: indestructible diplomats reporting on the decisions of a government at war, with only a QR code to prove that they are actually speaking for a government still lead by humans.

  1. Clément-Fontaine, Mélanie. “L’homo numericus.” (2019): 147-155. (https://hal.science/hal-02497174) ↩

  2. As an example of AI hallucinating and contradicting itself, I recently had a conversation of sorts with Chat GPT in which I asked it about some plot points in Star Trek Discovery. It got most of it right, but when I asked it who Mike Burnham was, it told me that there was no character named Mike Burnham in Star Trek Discovery – surely, I mean Michael Burnham. When I then asked who Michael Burnham’s biological father was, it told me her biological father was Mike Burnham, so I asked my previous question again, got an apology for the “confusion” and an explanation that there was no Mike Burnham character in Star Trek Discovery. When I asked about Michael Burnham’s biological father again, I was told his name was Sarek, and that he took her in when her (other?) biological parents died (the transcript of the chat, can be found here). ↩

  3. Even if we don’t believe that they have actually gotten anyone elected, or that they have directly corrupted politicians of various political hues, the West’s reaction to their known actions has undeniably affected policy decisions on both sides of the proverbial isle. ↩

  4. Nodoushan, Mohammad Ali Salmani. “The Quantum Human Computer (QHC) Hypothesis.” i-Manager’s Journal of Educational Technology 4, no. 4 (2008): 28. ↩

  5. Gao, Shan. “Are human brains quantum computers? Why quantum cognition says Yes.” (2019). ↩

  6. Mahmood, Tariq, and Rehan Ahmad Khan Sherwani. “Modeling of Quantum Cognitive Perceptual Associative Memory for Conscious Agents: A Parametrized Circuit of Quantum Neural Networks.” (2024) ↩

  7. Duan, Yucong. “Exploring Professor Yucong Duan’s Theory: A New Perspective from Quantum Consciousness to Artificial Intelligence Systems.” ↩

  8. Schroeder, Karl. “What If We Built a Really Big Computer?.” IEEE Spectrum 61, no. 3 (2024): 34-43. ↩