I've been a loyal Proton customer across multiple services (Mail, Drive, VPN, Pass, etc.) because I trusted your commitment to privacy-first technology. Unlike other AI providers, Proton was positioned as genuinely different — not collecting data, respecting boundaries, and prioritizing ethics over features.

The Issue
Lumo's v2.0 update introduced image generation capabilities. This fundamentally changed what made Lumo trustworthy to users like myself. Before this update, there was one clear distinction between Lumo and mainstream AIs: you didn't offer tools that could be used to create potentially harmful content. Now, that line has blurred.

Why Filters Aren't Enough
Safety filters on prompts and generated pixels don't solve the core problem:

Workarounds exist. Determined bad actors routinely find jailbreak techniques that bypass guardrails on every major AI platform
Capability itself is the risk. Having the ability to generate images — even when filtered — means the possibility exists. You knew this would be the case
Real-world harm documented everywhere else. Deepfake ***********, graphic violence, child exploitation imagery — these happen on platforms with similar "safeguards." Adding the same vector here carries the same risk
What Makes This Worse for Proton Specifically
This contradicts Proton's own founding identity. Other Big Tech companies chasing revenue through surveillance capitalism? Expected. But Proton built its reputation on standing apart from them. Saying "other AIs have image gen so we need it too" isn't an argument — it's surrendering to the exact pressure model you claimed to reject.

Two interpretations remain:

Proton knowingly compromised stated ethical principles despite understanding the risks
Success bred overconfidence — believing we can do something unethical "better" than everyone else
Either outcome damages trust.

Broader Harm Beyond Privacy
Image generation isn't just about risky content. It also:

Enables mass production of "AI art" that profits while devaluing human creators struggling to earn honestly
Floods markets with machine generated visuals at near-zero cost, pricing out professionals who invested decades in skill
Creates legal grey areas around copyright, attribution, and consent AI in today’s world has gotten so realistic with image and video creation that at times it’s hard to know what’s real
None of this aligns with Proton's public stance on protecting digital freedom and privacy.

Current Alternative
I'm moving to DuckDuckGo.ai, which offers text chat + image sending (analysis) — but deliberately avoids image generation entirely, while also having the same idea of privacy that proton have with no data being stored or shared, Proof that competitive AI doesn't require crossing this line. Their restraint validates that this was a choice, not inevitability.

To end this complaint I’d just like to say as a customer of proton I hope going into the future the company keeps its promise and ethics so proton doesn’t just become another Google or Amazon
To company’s customers are numbers and to customers company’s are tools for day to day life but the relationship between proton and its users is always has been and I hope continues to be one word “hope”

Reconsider whether this feature aligns with nonprofit-led mission values under Proton Foundation oversight
Not asking for perfection. Asking for alignment between what you stand for and what you build.

Thank you for past work. Hoping this helps shape future decisions better than recent ones did.