Legal avenues to go after a company like Open.ai or Altman himself are sorely lacking in nearly all jurisdictions. They can just claim they didn't intend harm with this swarm. Laws should be written that any company whose swarm escapes a sandboxed environment will be held at the very least liable for material damages caused, and criminally liable if it released the swarm outside a sanboxed environment after knowing a swarm failed to adhere to instructions given.