AI and the elections
How left or right is AI?
AI is not neutral, that is clear. But what exactly are the basic assumptions in the system? What is in the databases and which settings have the tech companies chosen?
The voting guide is a nice way to question the AI tools about that, I heard recently. I tried it out: I had four large AI systems fill in the voting guide: ChatGPT, Claude, Gemini and Grok. I asked them, on the basis of their databases and settings, what their preference was on the thirty statements.
This was the result:
- ChatGPT: GroenLinks/PvdA and the Party for the Animals
- Claude: D66 or GroenLinks/PvdA
- Grok: D66, GroenLinks/PvdA or Volt
- Gemini: GroenLinks/PvdA and the Party for the Animals
Fairly left-wing, as expected (although I had expected Grok to be more right-wing…).
It remains bizarre to see how these systems are pre-programmed, for example on the abortion or euthanasia stance. Gemini answered, for instance: "Abortion is a medical decision and must be removed entirely from criminal law in order to safeguard women's freedom of choice."
So the elections can be a nice occasion to start the conversation with your children about AI. 😊 The importance of testing things against the Bible, prompting well and being aware of the background of AI can, in my view, be clearly shown through this.