Your Open Source Model Could Have a Hidden Time-Release Backdoor

Tiberium

There is quite old research on this: - https://arxiv.org/abs/2311.14455 - https://arxiv.org/abs/2401.05566 - https://arxiv.org/abs/2410.13722

zarzavat

Yes your Chinese open model could have a time-release backdoor, just as your Chinese vibrator could have a hidden microphone that records everything you say and transmits it to the CCP. But does it? No. What's much more likely is that your US AI provider is promising not to train on your data but is doing so anyway. With a self-hosted model you can at least avoid that.

OutOfHere

We should subject all models to a temporaral stability benchmark looking forward up to a hundred years. Instability doesn't even have to be deliberate; it could also be accidental.