3 ms·
And how are you going to distinguish those interactions from chatbots trying to sell you something?
by sebzim4500 4y ago
And how are you going to distinguish those interactions from chatbots trying to sell you something?
- CuriouslyC 4y agoA network of trust, backed by a social graph, which can be used to filter untrusted content.
- sebzim4500 4y agoWhat if people start trusting the AI more than other people? It will tell them exactly what they want to hear.
- CuriouslyC 4y agoAI content will be associated with a user or organization in the trust graph. If someone you trust trusts a user or organization who posts AI content, you're free to revoke your trust in that person or blacklist the specific users/organizations you don't want to see anymore.
- pessimizer 4y agoWe've been pretending to be just about to do this for decades. The fact is that internet companies will not develop a network of trust, because they are primarily advertisers looking for better ways to abuse trust.
- chatmasta 4y agoOpenAI at least can track the hashes of all content it's ever output, and filter that content out of future training data. Of course they won't be able to do this for the output of other LLMs, but maybe we'll see something like a federated bloom index or something. Agreed there is no perfect solution though, and it will definitely be a problem finding high quality training data in the future.