Right to Intelligence: Campaign Seeks Legal Cover for Running Local AI Models

A campaign for local models

Right to Intelligence is a new advocacy campaign built around a single claim: local AI is the next personal computer. Its site contrasts a model that runs on your own machine — one you can inspect, repair and improve — with a chatbot account or a rented API, which depends on a platform staying online. The line the campaign says it exists to protect is the ability to run such models without asking permission.

Protect lawful use, enforce real harm

The campaign's stated position is that people should be free to download, own, run, study, modify and share open AI models. It also lists the harms it wants enforced seriously: fraud, cybercrime, child sexual abuse material, harassment, non-consensual intimate deepfakes, discrimination and sabotage. What it draws as the red line is narrower — requiring a licence merely to own or run the tool.

Local first, not cloud only

The argument for running models locally is practical. Not every AI task needs a warehouse of GPUs, the site says; many everyday workloads can run on a laptop, desktop or phone people already own. The campaign concedes that local AI will not replace every cloud model or training run, and asks only that the law not push simple, lawful tasks back into the cloud when the device can handle them.

Call scripts, state by state

The action the campaign asks for is state-level contact with legislators: choosing a state produces a call script, and the site says the state alone is enough to start. An official lookup may ask for an address to identify exact representatives, which the campaign says it does not store. Signing up takes an email address and a state. Its counters currently report 1,511 US signatures, 117 calls logged and activity across all 50 states.