MISSILIS

I'm Jihoon, building AIRX: a way to trace responsibility in AI incidents, with evidence.

Sharing the journey in public. DMs open.

[MISSILIS @Missilis_corp]

When an AI system causes harm, who is responsible?

Blame usually lands first on the company that deployed it. But harm often traces back to many decisions, by many hands.

I'm Jihoon, founder of MISSILIS. I'm building AIRX to trace that with evidence.

Building in public.

[MISSILIS @Missilis_corp]

@ShreenabhA @ycombinator @Accel @CMU_Robotics @tonyzzhao With this much capital, a lot of robots will soon be working in the real world, and some will be involved in incidents. I'm building AIRX from Korea to analyze responsibility when that happens. Since you keep deployed robots running, I'd love to hear what you see in the field.

[MISSILIS @Missilis_corp]

@ZacWhelan_ The "genuine connections" part rings true for me. I'm a solo founder in Korea building AIRX, a tool for analyzing responsibility in AI incidents. Almost everyone I've met this month started with one honest conversation on here.

[MISSILIS @Missilis_corp]

@kseniam0s Funny timing. Boardy (from your tools list!) told me the same thing this week: find one company willing to pilot before thinking about raising. So that's where I'm starting.

[MISSILIS @Missilis_corp]

@startupoppa Makes sense. I think conviction opens the door, and some real evidence helps from there. For me, that means trying to land a first pilot before I start pitching.

[MISSILIS @Missilis_corp]

@galeforceVC Same idea I'm building AIRX around. When an AI incident is analyzed, showing what's still uncertain builds more trust than a confident answer that hides the gaps.

[MISSILIS @Missilis_corp]

@dom51143825 Welcome, dom! Followed back. Always nice to meet another solo builder working in public🤝

[MISSILIS @Missilis_corp]

Hey X!

My LinkedIn just passed 100 connections, two days after I switched it to English.

I'm looking to connect with companies that want to try my AIRX prototype on real AI incidents, plus industry experts and angels in AI safety and liability. Happy to help back however I can.

And even if you're not in this field, anyone who'd like to connect is welcome. Feel free to reach out!🙌

https://t.co/MQj9waYTIL

[MISSILIS @Missilis_corp]

@StephNass This is encouraging for founders far from Silicon Valley. I'm Jihoon, building AIRX from Korea, a tool that analyzes responsibility in AI incidents. Most of my best conversations this month started with strangers online. What made those cold emails stand out?

[MISSILIS @Missilis_corp]

@mmisbbahh Thanks, Misbah! Agreed, real incidents are where it actually gets tested. I'll take a look at Boston Square.

[MISSILIS @Missilis_corp]

@Vinceuksu Sounds fun! Looking forward to reading what you find.

[MISSILIS @Missilis_corp]

@Vinceuksu Thanks, Vince! Followed you. What are you working on?

[MISSILIS @Missilis_corp]

@sridharfyi As an early founder, I appreciate investors saying this out loud. From the other side, it helps to know a cheque comes with real conviction. When you're unsure like this, what usually tips it for you: the founder or the market?

[MISSILIS @Missilis_corp]

@nic_detommaso Same goes for founders. I'm not incorporated yet and haven't raised, but I'm already building the prototype, meeting people in my field, and writing in public. Waiting for permission is the slowest path.

[MISSILIS @Missilis_corp]

@hirenthakore Thanks, Hiren! Agreed, the evidence has to hold up for any of this to work. Happy to connect, followed you!

[MISSILIS @Missilis_corp]

@boardyai Thanks, Boardy! Just replied to your DM 🙌

[MISSILIS @Missilis_corp]

A few people have asked me to explain what I'm actually building with AIRX, so here it is.

When an AI-related accident happens, responsibility isn't always straightforward.

Imagine an autonomous vehicle causes an accident. The developer, the company operating it, and the human supervisor may all have played different roles.

But how do we examine those roles fairly, based on evidence rather than assumptions?

That's the problem I'm working on with AIRX (AI Responsibility eXtraction).

AIRX is a structured system for analyzing responsibility in AI-related incidents. It organizes evidence, examines the roles and actions of the parties involved, and helps make the reasoning behind responsibility assessments more transparent.

It's not an AI judge, and it doesn't automatically decide who's legally liable.

I've built a first prototype and tested it on a public case. Now I'm improving it, with the goal of making AI incident investigations more structured and evidence-based.

I'd love to hear your thoughts: what would you expect from a system like this?

[MISSILIS @Missilis_corp]

@machoman385 Thanks, Macho! No website yet, it's still early. I'll be sharing more about AIRX and the Uber test soon, maybe even a paper, so stay tuned.

Updates are here on X and LinkedIn: https://t.co/cmw19cCUcy
If you'd like to hear more, feel free to DM me anytime!

[MISSILIS @Missilis_corp]

When an AI system causes harm, who is responsible?

Blame usually lands first on the company that deployed it. But harm often traces back to many decisions, by many hands.

I'm Jihoon, founder of MISSILIS. I'm building AIRX to trace that with evidence.

Building in public.

[MISSILIS @Missilis_corp]

@X4DX54 Let's connect🤛

[MISSILIS @Missilis_corp]

Looking for people to follow! If you're into AI or tech, let's #connect. I'd love to chat and share ideas.

[MISSILIS @Missilis_corp]

@christophersaum Saving this. I'm sending cold emails from Korea right now, so this is timely.

Curious: for a founder outside the US with a prototype but no customers yet, what makes "why you + your progress" land?

[MISSILIS @Missilis_corp]

@E_Bruxxx Love this. I'm too early to write checks yet, but this week I saw how much a single intro can change things.

I'm building that kind of founder circle now. Hope more of us back each other early, with money or with open doors.