Show HN: Witness – Livestreaming for Emergencies | Hacker News Reader

Show HN: Witness – Livestreaming for Emergencies

getwitness.com

57 pointsmarinosbern55 comments

Hi HN! I’m Marinos, I made this. Here’s a medium post I just wrote going into more detail on the motivation & design decisions behind Witness: https://medium.com/@marinosbern/witness-livestreaming-for-em.... Let me know if you have any feedback/questions about Witness or live streaming or anything else. One question I get a lot is whether I am planning to open source this. I am, and I’m also extracting all the live streaming logic into a drop-in iOS library. Cheers!

Awesome to see companies open source their tools.

I'm just one guy, not really a company:) But definitely agree with you, and, in fact, Witness itself is made possible thanks, in part, to a bunch of open source projects

After you're done polishing/publicizing your app, you should write a blog post describing the transition from developing trading algorithms to developing public service/safety apps :)

:) I was lucky enough to work for a great company [1] that was very techy so in terms of the culture it wasn't much of a transition. The hardest part (which I am still working on) is developing for anyone out there instead of internal use. I rarely had to think about UI/UX/User Behavior previously.

[1] http://www.sig.com/

Reminds me a bit of this: http://www.sosbeacon.org/ - I worked on an early version of it many years ago. Doesn't look like he's done much with it lately, though.

It's a good idea, so good luck!

This looks really nice - are you planning to support other OS, like Android, WP or Blackberry?

Blackberry? Seriously?

What? The Android apps are fairly easily ported.

Yes, I'm already working on an Android version and an Apple Watch version. I'm really excited about the new Apple watch which is rumored to have a camera because it seems like the perfect medium for Witness to run on

I am curious, did you do a search for the term and other name collisions before you started your project?

From http://witness.org/ : "WITNESS trains and supports activists and citizens around the world to use video safely, ethically, and effectively to expose human rights abuse and fight for human rights change."

They are also on github: https://github.com/witness

Unfortunately not - I picked the name at 4am at the Disrupt hackathon and, at that point, name collisions were the least of my worries. In fact, there's a ton of apps in the App Store with Witness in them. After it won the hackathon the name stuck and it was impossible to change

Awesome. I really like it.

An initial thought...Perhaps, it's use case is possibly a little more Western World-centric due to the audio/video aspect of things. Where recording a perpetrator is a deterrence against further aggression.

We do a lot of work with NGOs and human rights defenders where security, especially in developing nations, is a different concept. What I mean is that in many emergency situations, the last thing you want a normal person (starting to panic and who's decision making process is starting to slow down and become impaired under pressure) to do is start holding a video camera up. Especially in the context of a case like a kidnapping or under fire. That will only have only the effective of a) probably antagonising the attacker, leading to the phone being snatched and/or thrown away (thus breaking the most important aspect, the real time reporting of current location)...b) in the most high risk situations, the user needs to be getting themselves to safety, as this initial 30 second reaction is the most important - focusing on the video and audio on the phone will slow this down and undo the natural inbuilt reaction to run like hell...

Anyway, just a thought, awesome project and good on you for doing it.

(Please excuse the following: Plug for our security planning tool, which aims at preventing physical and digital security incidents http://www.secfirst.org ...Plug for a very cool Android panic button: Amnesty https://panicbutton.io)...Plugs over :)

I'm wondering how I would undo the "natural inbuilt reaction to run like hell" long enough to tap tap and look at my phone.

I actually asked a bunch of psychologists about this. Most people will either freeze or run. If your inbuilt reaction is to run, you are already in the lucky group. The goal is to use it, if possible, while running, so I tried to make it as simple as possible to activate. Unfortunately, because of Apple restrictions, you still need to launch the app, but I'm experimenting with things like using notification center for faster access

Cool. I'm not sure if your part of the community already but if it's open source etc you should post it on the Stamford University Liberation Tech mailing list.

Talk about an edge case for UI design :). I watched a show about a cop who couldn't open his trunk with his keys while he was under fire.

The psychology of people in high pressure, threat to life situations, is fascinating, a huge chunk of people freeze or do the most ridiculous stuff.

Well worth reading "The Unthinkable" by Amanda Riley, for an interesting overview. Though there is tons of academic literature on the subject.

I'll check it out, thanks for recommending

Some feedback from a discussion I had after seeing this:

three modes: 1 - immediate life threatening danger, call police (and notify friends/family) 2 - non-immediate, but still notify friends/family 3 - possible danger (just scared) - record everything (remote server), hold for 24 hours before sending to friends/family.

I think #3 could be one of your best use cases: A lot of people don't like walking to their car alone at night, or through a dark alley, or meeting someone on a first date. But they also don't want to stream that to their friends/family for no real reason. This gives them time to stop/erase content where nothing happens. In the case where something does happen, evidence is still maintained even if the phone is knocked/dropped out of their hand immediately, or whatever. You could also have a quicker way to "upgrade" to #1 in this mode, in case the threat turns more immediate and the user still has time/thought process to act.

You've hit the nail right on the head. A 2-stage process (1-record 2-notify) is definitely the number one feature request. For this version, I wanted to keep things super simple and get it out there as soon as possible so I can start getting feedback. I should be adding that to the next version.

Would you prefer an automatic 24-hour notification or a manual 2-stage process where you first record and then notify the emergency contacts if needed?

Gotcha. Good on you for getting out a simple version quickly, that's the way to do it.

Personally I prefer automatic notification, because the user may be impaired and can no longer perform the second step.

However, I would suggest you make this all user customizable. Length of time before automatic notification, automatic vs manual, etc.

Congratulations! A similar app won Google's Social Impact Challenge 2014 [1] in Brazil. It's named PLP 2.0 [2], and it's aimed mainly at women that suffer from violent partners.

A few friend and I built something similar for a hackathon made by the brazilian Chamber of Deputies on Genre [3]. Our main differentiator was that we didn't rely on having a smartphone at all, as our focus was on low-income women in Brazil, which (we thought) is the class that needs something like this the most. Although smartphone use is increasing very quickly in Brazil, only 20% of the brazilians had one in 2014.

Ours worked by having a phone number where people would call via a fast-dial number (so it would need just one button press), then it would start recording the audio and send an SMS to NGOs that work with women's protection so they could call the cops or whatever was deemed necessary.

Obviously, by not requiring a smartphone we wouldn't have access to the person's location. The solution was based on the fact that ~70% of the attacks occur on the victim's home. We would pre-register the person's number and addresses (home, work, etc.), so that when we received a "help request", we would know where to start looking for her. This was the best we could do with a dumb phone.

[1] https://desafiosocial.withgoogle.com/brazil2014 [2] http://www.plp20.org.br/ [3] https://github.com/thacker/botao-do-panico

Most phones have really terrible UIs for getting to something like this. In circumstances where I want to start using an app like this, I usually don't have 5-10 seconds to take my concentration off of my surroundings to load the app. I've been in plenty of situations where I wished I could immediately start recording things but didn't have time to navigate my phone's UI.

My experience was exactly the same. The record button on the camera app is small and hard to hit if you're running/your hands are shaky. Also if you touch the screen while recording it changes the focus and you could also accidentally end the recording. All these could have serious implications during an emergency.

That's why for Witness I made sure you could touch anywhere on the screen while recording and it would be almost impossible to do any damage. So you can conceal your phone while your fingers rest on the screen.

But how do you get into the app? Don't you have to navigate your phone's OS? I can appreciate what you're saying. There are probably lots of situations where you know you might need it where you can have the app already running. I just wish there was a way to get into something like this or start recording without even having to look at my phone. I think it would be much more useful if you had that.

Not a new concept. Many have tried and failed. Even though the social cause is clear and many people will say great idea, very few will use it if their life is on the line. And even less will pay for it unless it's bundled with some sort of home care service/monitoring.

The 1st point is actually a big liability problem. If your life is at risk will you make a 911 call (which even works without data coverage) or trust an app that you have not started in a while to work as expected? And you must have data for it to work too right. Apps need consistent updates to fix bugs and when a OS update breaks it, you are putting people at risk. Too much of a risk to take to rely on an app to save your life.

The 2nd point, is whether you will end up having a business. You have a business when someone pays for your product/service. Until then it is not a business. Who will pay for this luxury of a better 911 and why do they really need it when 911 is free and more reliable? If the emergency is not 911 serious level than what role does the app play?

Hope these questions help you flesh your thinking around this space/product. Just because you build it don't mean they will come ;-)

I love this concept, and it is sorely needed. I read in your Medium post[1] that “Data is wiped from the phone as soon as it’s securely moved to the Witness servers.” I'm curious what the reasoning for this is. Of course, I get that the data is streamed/copied to Witness in the event that the device is destroyed or confiscated, but it's hard for me to think of a use case where having a local version on the device would would do any harm.

In fact, I think it would be valuable for evidentiary purposes to have the original on the device. I'm assuming that the streamed/copied version is probably lossy in some regard, while the local version might be higher resolution or frame rate, also.

[1] https://medium.com/@marinosbern/witness-livestreaming-for-em...

> but it's hard for me to think of a use case where having a local version on the device would would do any harm.

Hostile environments. It becomes incriminating evidence against you (think recording police brutality or excessive uses of force).

I just watched this documentary "Point and Shoot" [1], the guy filmed all over Africa, including Libya, and at one point was taken prisoner where he said his footage could have been used to identify other members of the Libyan resistance. It isn't hard for me to imagine a situation where you don't want to give the authorities visual evidence of who was at an event or who you were standing with.

1: https://en.wikipedia.org/wiki/Point_and_Shoot_(film)

Thanks for pointing this out. The original thinking was twofold: 1) I didn't want people to get in trouble if their phones were searched and they were in possession of a recording and 2) Keep the data stored on the phone to a minimum, in order to increase available space for storing video when Cellular Data/WiFi is not available. The data is always available later through the shortlink.

You raise a very good point though. Offering an option to keep data locally (or save it in the camera roll) in a future version would make sense

Might be worth clarifying that the project is separate from the work of the well known human rights and video NGO called "Witness" (www.witness.org). They make a number of video tools so a user might be confused.

At the border US customs ask you whether you have "accounts on remote computers", which means, practically everything. It's better to move the data offsite so police actually needs a warrant to inspect it (instead of a "I thought he had a gun" excuse), but it can still require them to provide access to the remote account to the police.

Interesting. Have you been asked that before? I've done a fair amount of going in and out of the country and was never asked

Cool work...I haven't downloaded the app to my phone yet, but is there a stage between activating the app and then going into emergency mode, e.g. broadcast + emailing all emergency contacts?

I could foresee situations in which someone is fearful, but not yet ready to raise the alarm...in which case, having the phone go into video/audio/geo-broadcast mode would be nice, but not "email/text everyone".

I really liked the accompanying blog post (https://medium.com/@marinosbern/witness-livestreaming-for-em...)...a lot of nice meditations on making something new by noticing the small inconveniences and barriers in real-world usage of existing solutions.

I completely agree, having a 2-stage process to first stream and then notify emergency contacts is one of the top feature requests. I should be adding that in the next version

This is an excellent idea. I wish you the best of luck.

I have experience with social networking apps and a BIG challenge that you'll have is if the functionality depends on the person's friends onboarding for the person to get value.

I would encourage you to think very hard about providing value for the user without any of their friends necessarily being on the app. Even if you had a 90% completion rate for each step of your on boarding funnel, the longer you make your funnel, the more you'll trend toward 0% activation.

Thanks! The functionality does not depend on the person's friends joining the app

Great idea. What happens if there's no cell network? Does it record AV and upload later in the background?

If there's no cell network it will store everything locally (status bar turns purple) until the cell network comes back up. This happens automatically and I'm already working on making it more explicit in future versions

Good stuff, thanks for the update.

Awesome! saw it on techcrunch: http://techcrunch.com/2015/05/03/witnesss-livestreaming-app-...

I like the idea of combining this with drones: http://www.civileyes.me/

Great work! I've heard of this for Uber in India, but always thought that a standalone version would be a great idea. Genius all around.

Yes, I have heard about this as well. A bunch of people from India have emailed me about using Witness there and some have already been field-testing it. This version does work in India, with full calling/texting support, but I definitely want to work on it a little more to better tailor it to places outside the US

The link to the privacy policy just links to the promotion webseite. Do you plan on selling footage to news organizations?

I think this goes without saying, but absolutely not. Data will never be available to anyone in any way, shape or form, except people chosen by you. I am adding this to the website now.

Very good idea! But what happens if something's happening in underground parking lot where there's no signal?

It will store everything on the device until the Internet is back up, then it will resume the stream. If it never comes back up, it will keep storing on the device until battery/storage runs out. It is designed to never lose data unless it has to

I've witnessed this app at techCrunch, such a simple yet so powerful tool for safety.

The title and event his web page led me to believe it was related to this project: http://witness.org/

"WITNESS is an international organization that trains and supports people using video in their fight for human rights."

I am so sorry for that, I picked the name Witness at 4am during a hackathon and had no idea it would win. It was an unfortunate choice of name, especially given that there are a ton of apps in the App Store with "Witness" in them

No, I don't think it's a bad thing! I was actually thinking about how the areas of overlap could be expanded and connected.

Was this inspired by the last 2 episodes of Silicon Valley?

Haha I've never seen Silicon Valley but I will sure check it out!

Reply on news.ycombinator.com