The same should apply to "Smart Contracts" in my opinion. On this topic myself and Vlad Zamfir agree. It makes sense from an ethical and security perspective that a community of some sort always have the capability to 1) change the utility function of any smart contract at any given time by means of consensus or 2) ability to shut down, cancel or stop any process or smart contract which uses the resources of the community. This may in fact need to be seen as a right, the right of the community to shut down any agent in the multi-agent system.

The reason Ethereum required a hard fork was because Ethereum did not have the proper governance and disaster recovery mechanisms set up. In corporate cybersecurity disaster recovery is pretty much mandatory procedure.  Yet when it comes to smart contracts there aren't yet good tools for handling these situations and The DAO represented an early manifestation of that. In my opinion more attention will need to be paid to security and ethics which means we need philosophical discussions as well as technical security discussions when in the whitepaper stage of a project. It may even be beneficial to come up with some community standards for secure and ethical smart contracts which can give a guideline or recommendations to act as a sort of "best practices".