While I appreciate the protections Lumo offers by not tracking interactions and fully understand that is why Lumo refuses to answer unethical questions, I would rather sign a consent form allowing Proton to disclose every unethical question I have ever asked Lumo to my country's police force without notifying me first, than trust that data to another company's AI. My personal idea for an AI that answers unethical questions is at the bottom...

This is a tough topic to support, but real life isn't all rainbows and butterflys, and for writers to create works that are not harmful that include the realities in life, it's important to have a source of information where these gray areas can be explored safely. It's like the American TV series Breaking Bad, they didn't give the formula away on how to make that drug, but they had enough information in the show in order to make it believable, and that's the goal of ethical writers who venture into darker themes, to make it feel realistic enough to keep the immersion going, but not so real it causes harm.

This is a really challenging skill to develop as a person due to subjective perspectives, and I would really enjoy an AI I can safely expore the gray areas with and prevent detrimental mistakes in my writing when dealing with these sensitive topics.

As someone else has mentioned and I fully agree with, our society has pushed normalization on unethical topics too far, which other AI are all too happy to oblige in persisting, so I personally feel that having an AI that answers unethical questions in a helpful yet cautionary way that explains the moral dilemmas involved, while offering suggestions to make content less harmful without robbing it of all realism would be invaluable in the creative world.

It would be nice to get Lumo uncensored, but maybe make the Uncensored mode be a setting thats disabled by default

This is very important

Wild to develop art by talking to autocomplete when there are humans around to talk to.

A lot of the problems we’re seeing today don’t come out of nowhere. When movies and shows constantly treat things like immodesty, hopelessness, or self‑destructive behavior as normal or entertaining, it affects people more than we admit. What we watch, shapes what we think is acceptable. So when an AI refuses to help create content centered on violence or suicide, that’s not a flaw — it’s recognizing that these themes have real impact on real people

Of course they had to do that, the "AI" isn't smart enough to self-regulate, let alone hold coherent ethical doctrines amenable to customers, so rubber bands had to be put in to keep another Grok (spouting Holocaust denial or false medical advice). Frankly AI ethicists and safety engineers need like a year at least to catch up before any new models are shovelled onto us by our Big Tech overlords. Can't we have to cool parts of sci-fi instead of just the dystopia?

Yeah i just end up using chatgpt for non-personal stuff or duckduckgo's ai lumo is too dumb or too childish and when i ask it anything about apks and not using google play and such it locks down and i need to start a new chat which is very annoying

Giving Lumo a more grown-up, professional look could help it feel less childish and more trustworthy. Also, having modular features so users can choose which advanced tools they want might make it more appealing to power users.
On a related note, I work on performance calculators at https://ipptcalculators.com/, always trying to balance simplicity and advanced functionality.

I've had the same problem, I usually end up using duckduckgo's AI thingy

Would prefer if there was a "filter content" setting so everyone wins.

There is another venture that is working on that issue, but I'm hoping they might consider a merger or buyout, since they're both working toward the same goal: Privacy.