Some 40 years ago, the sci-fi/horror film Demon Seed told the tale of a woman slowly imprisoned by a sentient AI, which invaded the smart home system her husband had designed to manage it. The AI locked doors, windows, turned off communications, and even put a synthesised version of her onscreen at the front door to reassure visitors she was “fine.”

The reality, of course, is that she was anything but. There’s been endless works of fiction where smart technology micromanaging the home environment have gone rogue. Sadly, those works of fiction are bleeding over into reality.

In 2018, we suddenly have the real-world equivalent playing out in homes and behind closed doors. We’ll talk about the present day problems momentarily, but first let’s take a look how we got here by casting our eye back about 15 years ago.

PC spyware and password theft

For years, a subset of abusive partners with technical know-how have placed spyware on computers or mobile devices, stolen passwords, and generally kept tabs on their other half. This could often lead to violence, and as a result, many strategies for defending against this have been drawn up down the years. I effectively became involved in security due to a tech-related abuse case, and I’ve given many talks on this subject dating back to 2006 alongside representatives from NNEDV (National Network to End Domestic Violence).

Consumer spyware is a huge problem, and tech giants such as Google are funding programs designed to help abused spouses out of technological abuse scenarios.

The mobile wave and social control

After PC-based spyware became a tool of the trade for abusers, there came an upswing in “coercive control,” the act of demanding to check emails, texts, direct messages and more sent to mobile phones. Abusive partners demanding to see SMS messages has always been a thing, but taking your entire online existence and dumping it into a pocket-sized device was always going to raise the stakes for people up to no good.

Coercive control is such a serious problem that the UK has specific laws against it, with the act becoming a crime in 2015. Should you be found guilty, you can expect to find yourself looking at a maximum of five years imprisonment, or a fine, or both in the worst cases. From the description of coercive control:

Coercive or controlling behaviour does not relate to a single incident, it is a purposeful pattern of incidents that occur over time in order for one individual to exert power, control, or coercion over another.

Keep the “purposeful pattern of incidents occurring over time in order for an individual to exert power or control” description in mind as we move on to the next section about Internet of Things (IoT) abuse, because it’s relevant.

Internet of Things: total control

An Internet of Things control hub could be a complex remote cloud service powering a multitude of devices, but for most people, it’s a device that sits in the home and helps to power and control appliances and other systems, typically with some level of Internet access and the possibility of additional control via smartphone. It could just be in charge of security cameras or motion sensors, or it might be the total package: heating and cooling, lighting, windows, door locks, fire alarms, ovens, water temperature—pretty much anything you can think of.

It hasn’t taken long for abusive partners to take advantage of this newly-embedded functionality, with numerous tales of them making life miserable for their loved ones, effectively trapped in a 24/7 reworking of a sci-fi dystopian home.

Their cruelty is only limited by what they can’t hook into the overall network. Locking the spouse into their place of residence then cranking up the heat, blasting them with cold, flicking lights on and off, disabling services, recording conversations, triggering loud security alarms; the abused partner is almost entirely at their mercy.

There are all sorts of weird implications thrown up by this sort of real-world abuse of technologies and individuals. What happens if someone has an adverse reaction to severe temperature change? An epileptic fit due to rapidly flickering lights? How about someone turning off smoke alarms or emergency police response technology and then the place burns down or someone breaks in?

Someone could well be responsible for a death, but how would law enforcement figure it out, much less know where to pin the blame?

Of course, those are situations where spouses are still living together. There are also scenarios where the couple has separated, but the abuser still has access to the IoT tech,  and they proceed to mess with their lives remotely. One is a somewhat more straightforward to approach than the other, but neither are particularly great for the person on the receiving end.

A daunting challenge

Unfortunately, this is a tough nut to crack. Generally speaking, advice given to survivors of domestic abuse tends to err on the side of extreme caution, because if the abuser notices the slightest irregularity, they’ll seek retribution. With computers and more “traditional” forms of tech-based skullduggery, there are usually a few slices of wiggle room.

For example, an abused partner may have a mobile device, which is immediately out of reach from the abuser the moment they go outside—assuming they haven’t tampered with it. On desktop, Incognito mode browsing is useful, as are domestic abuse websites which offer tips and fast close buttons in case the abuser happens to be nearby.

Even then, though, there’s risk: the abuser may keep network logs or use surveillance software, and attempts to “hide” the browsing data may raise suspicions. In fact, this is one example where websites slowly moving to HTTPs is beneficial, because an abuser can’t see the website data. Even so, they may still see the URLs and then you’re back to square one.

With IoT, everything is considerably much more difficult in domestic abuse situations.

A lot of IoT tech is incredibly insecure because functionality is where it’s at; security, not so much. That’s why you see so many stories about webcams beamed across the Internet, or toys doing weird things, or the occasional Internet-connected toaster going rogue.

The main hubs powering everything in the home tend to be pretty locked down by comparison, especially if they’re a name brand like Alexa or Nest.

In these situations, the more locked down the device, the more difficult it is to suggest evasion solutions for people under threat. They can hardly jump in and start secretly tampering with the technology without notice—frankly people tend to become aware if a physical device isn’t acting how it should a lot faster than their covert piece of spyware designed to grab emails from a laptop.

All sorts of weird things can go wrong with some purchased spyware. Maybe there’s a server it needs to phone home to, but the server’s temporarily offline or has been shut down. Perhaps the Internet connection is a bit flaky, and it isn’t sending data back to base. What if the coder wasn’t good and something randomly started to fall apart? There’s so many variables involved that a lot of abusers might not know what to do about it.

However, a standard bit of off-the-shelf IoT kit is expected to function in a certain way, and when it suddenly doesn’t? The abuser is going to know about it.

Tackling the problem

Despite the challenges, there are some things we can do to at least gain a foothold against domestic attackers.

1) Keep a record: with the standard caveat that doing action X may attract attention Y, a log is a mainstay of abuse cases. Pretty much everyone who’s experienced this abuse and talks about it publicly will say the same thing: be mindful of how obvious your record is. A book may work for some, text obfuscated in code may work for others (though it could attract unwarranted interest if discovered). It may be easier to hide a book than keep them away from your laptop.

Of course, adjust to the situation at hand; if you’re not living with the abusive partner anymore, they’re probably not reading your paper journal kept in a cupboard. How about a mobile app? There are tools where you can detail information that isn’t saved on the device via programs designed to look like weather apps. If you can build up a picture of every time the heating becomes unbearable, or the lights go into overdrive, or alarms start buzzing, this is valuable data for law enforcement.

2) Correlation is a wonderful thing. Many of the most popular devices will keep detailed statistics of use. Nest, for example, “collects usage statistics of the device” (2.1, User Privacy) as referenced in this Black Hat paper [PDF]. If someone eventually goes to the police with their records, and law enforcement are able to obtain usage statistics for (say) extreme temperature fluctuations, or locked doors, or lightbulbs going berserk, then things quickly look problematic for the abuser.

This would especially be the case where device-recorded statistics match whatever you’ve written in your physical journal or saved to your secure mobile app.

3) This is a pretty new problem that’s come to light, and most of the discussions about it in tech circles are filled with tech people saying, “I had no idea this was a thing until now.” If there is a local shelter for abused spouses and you’re good with this area of tech/security/privacy, you may wish to pop in and see if there’s anything you could do to help pass on useful information. It’s likely they don’t have anyone on staff who can help with this particular case. The more we share with each other, the more we can support abused partners to overcome their situations.

4) If you’ve escaped an abusive spouse but you’ve brought tech with you, there’s no guarantee it hasn’t been utterly compromised. Did both of you have admin access to the devices? Have you changed the password(s) since moving? What kind of information is revealed in the admin console? Does it mention IP addresses used, perhaps geographical location, or maybe a new email address you used to set things up again? If you’ve been experiencing strange goings on in your home since plugging everything back in, and they resemble the type of trickery listed up above, it’s quite possible the abusive partner is still up to no good.

We’ve spotted at least one example where an org has performed an IoT scrub job. The idea of “ghosting” them, which is keeping at least one compromised device running to make the abuser think all is well is an interesting one, but potentially not without risk. If it’s at all possible, our advice is to trash all pieces of tech brought along for the ride. IoT is such a complex thing to set up, with so many moving parts, that it’s impossible to say for sure that everything has been technologically exorcised.

No quick fix

It’d be great if there was some sort of technological magic bullet that could fix this problem, but as you’ll see from digging around the “IoT scrub job” thread, a lot of security pros are only just starting to understand this type of digitized assault, as well as the best ways to go about combatting it. As with all things domestic abuse, caution is key, and we shouldn’t rush to give advice that could potentially put someone in greater danger. Frustratingly, a surprising number of the top results in search engines for help with these types of attack result in 404 error pages or websites that simply don’t exist anymore.

Clearly, we all need to up our game in technology circles and see what we can do to take this IoT-enabled horror show out of action before it spirals out of control. As IoT continues to integrate itself into people’s day-to-day existence, in ways that can’t easily be ripped out afterwards, the potential for massive harm to the most vulnerable members of society is staring us in the face. We absolutely must rise to the challenge.