One way of finding new links between genes and diseases is to sequence a patient’s exome. Photo: Getty
Show Hide image

DIY diagnosis: how an extreme athlete uncovered her genetic flaw

When Kim Goodsell discovered that she had two extremely rare genetic diseases, she taught herself genetics to help find out why. Ed Yong tells her story.

Kim Goodsell was running along a mountain trail when her left ankle began turning inward, unbidden. A few weeks later she started having trouble lifting her feet properly near the end of her runs, and her toes would scuff the ground. Her back started to ache, and then her joints too.

This was in 2002, and Kim, then 44 years old, was already an accomplished endurance athlete. She cycled, ran, climbed and skied through the Rockies for hours every day, and was a veteran of Ironman triathlons. She’d always been the strong one in her family. When she was four, she would let her teenage uncles stand on her stomach as a party trick. In high school, she was an accomplished gymnast and an ardent cyclist. By college, she was running the equivalent of a half marathon on most days. It wasn’t that she was much of a competitor, exactly – passing someone in a race felt more deflating than energising. Mostly Kim just wanted to be moving.

So when her limbs started glitching, she did what high-level athletes do, what she had always done: she pushed through. But in the summer of 2010, years of gradually worsening symptoms gave way to weeks of spectacular collapse. Kim was about to head to Lake Superior with her husband, CB. They planned to camp, kayak, and disappear from the world for as long as they could catch enough fish to eat. But in the days before their scheduled departure, she could not grip a pen or a fork, much less a paddle. Kim, a woman for whom extreme sports were everyday pursuits, could no longer cope with everyday pursuits. Instead of a lakeside tent, she found herself at the Mayo Clinic in Rochester, Minnesota.

After four days of tests, Kim’s neurologist told her that she had Charcot–Marie–Tooth disease, a genetic disorder that affects the peripheral neurons carrying signals between the spinal cord and the extremities. It’s rare and carries a varying suite of symptoms, but Kim’s are typical, starting at the feet and heading upward. The neurologist explained that as her neurons died, the surviving cells picked up the slack by sprouting new branches – a workaround that masked the underlying degeneration until the rate of cell death outpaced the rate of compensation. Hence Kim’s crash.

The neurologist told her to come back in a year so he could check how quickly the disease was progressing, but that it would certainly progress. Charcot–Marie–Tooth has no cure.

The Goodsells drove home and Kim, exhausted, slept for two days. When she woke up, she got to work. “My reaction to things that I have no control over is to find out as much as I can about them,” she says. She started by reviewing her clinic notes, and quickly noticed something odd: there was hardly any mention of her heart.

Years before she learned that she had Charcot–Marie–Tooth, Kim discovered that she had another genetic disorder – one that affects the heart, arrhythmogenic right ventricular cardiomyopathy (ARVC). ARVC gradually replaces the heart’s synchronised beating muscle with fat and scar tissue. It nearly killed her once; she still has an internal defibrillator to keep her heart beating. But even though it was there in her medical records, her neurologist hadn’t seen fit to mention it in his report. “It meant nothing to him,” says Kim. “I thought: Wow, that’s really funny.”

It wasn’t the omission per se that bothered her. It was the implicit suggestion that her two life-long diseases – one of the heart, one of the nervous system – were unrelated. That, in the genetic lottery, she was a double-loser. That lightning must have struck her twice.

Surely not, she thought. Surely there must be a connection.

§

I meet Kim at La Ventana in Baja California, Mexico. She spends winters here, mostly kitesurfing. The sand and water are postcard-quality, but La Ventana has barely any resorts or big hotels. So in the still air of the morning when kites won’t fly, the beach is empty. Kim likes it that way. She has been up since dawn, cycling among the cacti and swimming in the ocean with pelicans and frigatebirds for company. She hauls herself out of the water, dries off, and sits on a small terrace overlooking the ocean. Her face is tanned and wrinkled, and she manifests no obvious signs of her two conditions. That’s partly because she has developed workarounds to mask and control her symptoms. She brushes her teeth on one foot to offset her balance problems. She uses massage balls and spends hours stretching to stop her muscles and joints from seizing up.

“See how I’m sitting?” she says. She has pulled her legs up on the chair to her left, and her back is curving that way too.

“My spine curves this way” – she nods to the right – “so I sit curving to the opposite side. I consciously do the opposite.”

She has a history of that. In 1979 Kim was a mathematically gifted pre-med student at UC San Diego, her hometown college. Her path was clear: graduate, and follow her older brother into medical school. But on a trip to South America – her first time out of San Diego – she ended up hiking for three months instead of working at a clinic as she’d planned. When she returned home, her academic future seemed pale and uninspiring. And then CB – her future husband, at this point a fellow student and regular running partner – started taking her out on wilderness hikes. “He introduced me to the mountains and I thought: this is life,” Kim says.

Within months of graduating Kim dropped out. Her brother, who had been a father figure to her growing up, was furious. “We hardly spoke. CB was his friend and he couldn’t even look at him,” she says. “He said I was being completely irresponsible.” Kim and CB married in 1983, and aside from a brief stint as restaurant owners, they have never had 9-to-5 jobs. They mostly earned a living by buying and remodelling run-down houses and selling them at a profit, and then heading into the wilderness until their supplies ran out. In 1995 they found themselves in La Jolla, California, working on an especially stressful renovation that left Kim drained.

That was when her heart problems began. Kim started having episodes of ventricular tachycardia – the lower chambers of her heart contracted so quickly that they pumped out their contents before they had a chance to fill up, compromising the flow of blood (and therefore oxygen) to the rest of her body. One minute she would be racing down Highway 1 on her bike; the next she would feel like she had been “unplugged”, as if “there was nothing driving anymore”. A cardiologist at Scripps Memorial Hospital told her she’d need an internal defibrillator, but Kim said no – she was worried it’d get in the way of wearing a backpack on a run, and she had faith that she’d be able to deal with the ventricular tachycardia by slowing down and relaxing. “I didn’t want something implanted in me that would limit my opportunities of experiencing life,” she says.

The next week, the Goodsells finished their renovation, packed up and headed into the Sierra Nevada with no return date in sight. It was an unorthodox solution to a life-threatening heart condition: to vanish into the boondocks, far away from any medical care, to do even more exercise.

The thing is, it was the right one. The outdoors rejuvenated her. She was gone for one-and-a-half years, and her heart behaved the whole way through. That unbroken streak only broke when the Goodsells rejoined their old lives in 1997. Back in California, they were once again cycling down Highway 1 when her heart started to beat erratically again. This time, it did not stop.

By the time the paramedics arrived, Kim was slumped against a wall and her chest was shaking. Her tachycardia had lasted for almost an hour and progressed to ventricular fibrillation – that is, her heartbeat was erratic as well as fast. She blacked out in the ambulance, on the cusp of cardiac arrest.

She woke up at Scripps Memorial Hospital. The same cardiologist was there to greet her. Through further tests he discovered that the muscle of her right ventricle was marbled with fat and scar tissue and not contracting properly. These are classic signs of ARVC. It had only been properly described in 1982, back when Kim was regularly signing up for triathlons. ARVC is a major cause of fatal heart attacks in young people, and athletes are especially vulnerable as exercise can accelerate the disease’s progress. And since Kim wouldn’t stop exercising, she finally conceded to the defibrillator. They implanted it the next day.

Kim referred to the implant as her “internal terrorist”. Every shock was debilitating and led to months of anxiety. She had to learn to cope with the device, and it took several years to regain the joy she drew from hardcore exercise. That was when the other symptoms started.

§

These diseases are rare. In a crowd of a million adults, around 400 will have Charcot–Marie–Tooth and between 200 and 400 will have ARVC. But genetic diseases in general are actually quite common – 8 per cent of people have at least one. This paradoxical combination has fuelled the rise of many online communities where people with rare disorders can find each other. Heidi Rehm, a geneticist at Harvard Medical School, studies a condition called Norrie disease that mostly affects the eyes and ears. She developed a registry for Norrie disease patients to share their experiences, and learned that almost all the men with the disease had erectile dysfunction. “A patient goes to their doctor with blindness and deafness, and erectile dysfunction isn’t the first thing you ask about!” says Rehm. “Patients drove that discovery.” Through communities, families often make connections about their medical problems that their doctors miss.

But Kim was never one for relying on others. She tried a support group when she got her implant, but it did nothing for her. She dipped her toes in patient forums, but was always frustrated by the rampant misinformation. “People just weren’t interpreting things correctly,” Kim says. “I wanted more rigour.”

She started by diving into PubMed – an online search engine for biomedical papers – hunting down everything she could on Charcot–Marie–Tooth. She hoped that her brief fling with a scientific education would carry her through. But with pre-med knowledge that had been gathering dust for 30 years and no formal training in genetics, Kim quickly ran headfirst into a wall of unfamiliar concepts and impenetrable jargon. “It was like reading Chinese,” she says.

But she persisted. She scratched around in Google until she found uploaded PDFs of the articles she wanted. She would read an abstract and Google every word she didn’t understand. When those searches snowballed into even more jargon, she’d Google that too. The expanding tree of gibberish seemed infinite – apoptosis, phenotypic, desmosome – until, one day, it wasn’t. “You get a feeling for what’s being said,” Kim says. “Pretty soon you start to learn the language.”

“Kim has an incredible ability to understand the genetic literature,” says Martha Grogan, a cardiologist from the Mayo Clinic and an old friend of CB’s who now coordinates Kim’s care. “We have a lot of patients who ask great questions but with Kim, it’s like having another research fellow.”

At the time the Goodsells were staying at a friend’s house at Lake Michigan. Kim would sit on the balcony for eight hours a day, listening to the water and teaching herself genetics. Too weak to explore winding hillside trails, she channelled her perseverance and love of isolation towards scientific frontiers and the spiralling helices of her own DNA. “I spent hundreds of hours,” she says. “CB lost me during this process.”

Kim looked at every gene linked to Charcot–Marie–Tooth – there are more than 40 overall, each one imparting a slightly different character to the disease. One leapt out: LMNA, which codes for a group of rope-like proteins that mesh into a tangled network at the centre of our cells. This ‘nuclear lamina’ provides cells with structural support, and interacts with a bunch of other proteins to influence everything from the packaging and activation of genes to the suicide of damaged cells. Given this central role, it makes sense that mutations in LMNA are responsible for at least 15 different diseases, more than any other human gene. These laminopathies comprise a bafflingly diverse group – nerve disorders (like Charcot–Marie–Tooth), wasting diseases of fat and muscle, and even premature ageing.

As Kim read about these conditions and their symptoms, she saw her entire medical history reflected back at her – the contracted muscles in her neck and back, her slightly misaligned hips and the abnormal curve in her spine. She saw her Charcot–Marie–Tooth disease.

She also saw a heart disorder linked to the LMNA gene that wasn’t ARVC but which doctors sometimes mistake for it. “Everything was encapsulated,” she says. “It was like an umbrella over all of my phenotypes. I thought: This has to be the unifying principle.”

§

Kim was convinced that she had found the cause of her two diseases, but the only way to know for sure was to get the DNA of her LMNA gene sequenced to see if she had a mutation. First, she had to convince scientists that she was right. She started with Grogan, presenting her with the findings of her research. Grogan was impressed, but pragmatic. Even if Kim was right, it would not change her fate. Her implant was keeping her heart problems under control, and her Charcot–Marie–Tooth disease was incurable. He didn’t see a point. But Kim did. “I wanted to know,” she says. “Even if you have a terrible prognosis, the act of knowing assuages anxiety. There’s a sense of empowerment.”

In November 2010 Kim presented her case to Ralitza Gavrilova, a medical geneticist at the Mayo Clinic. She got a frosty reception. Gavrilova told Kim that her odds of being right were slim. “I got this sense that she thought I’d made an unfounded shot in the dark,” says Kim. “That I didn’t understand the complexity of the genome. That I had been reading the internet, and they come up with all sorts of things there.”

Gavrilova pushed Kim towards a different test, which would look at seven genes linked to ARVC. Her insurance would cover that, but if she insisted on sequencing the DNA of her LMNA gene, she would have to foot a $3,000 bill herself. Why waste the money, when it was such an unlikely call? But Kim was insistent. She knew that the known ARVC genes explain only a minority of cases and that none of them was linked to neural problems. In all her searching she had found only one that covered both her heart and nervous problem. Eventually, Gavrilova relented.

Kim, meanwhile, disappeared down to Baja in Mexico. Gavrilova’s scepticism had worn her down and she fully expected that the results would come back negative.

When she returned home in May, there was a letter waiting for her. It was from Gavrilova. She had been trying to call for months. The test had come back positive: on one of her two copies of LMNA Goodsell had a mutation, in a part of the gene that almost never changes. LMNA consists of 57,517 DNA ‘letters’, and in the vast majority of people (and most chimps, monkeys, mice and fish) the 1,044th position is filled by a G (guanine). Kim had a T (thymine). “All evidence suggests that the mutation found in this patient might be disease-causing,” Gavrilova wrote in her report.

In other words, Kim was right.

“I’m beyond impressed,” says Michael Ackerman, a geneticist at the Mayo Clinic. He specialises in inherited heart disorders like ARVC that can cause sudden death at any time. Such diseases make for people who do their homework, but Ackerman describes most as “Google-and-go” patients who check their diagnosis online, or read up about treatment options. Kim had written up her research as a white paper – 36 pages of research and analysis. “Kim’s the only one who handed me her own thesis,” he says. “Of all the 1,000-plus patients I’ve taken care of, none have done extensive detective work and told physicians which genetic test to order.”

He thinks she nailed it too. It is unlikely to be the whole story – Kim almost certainly has other mutations that are affecting the course of her disease – but LMNA “is certainly the leading contender for a unifying explanation, without there being a close second,” he says. “The evidence is pretty good for this being a smoking gun.”

The test had vindicated her hypothesis, but it also raised some confusing questions. Heart problems are a common feature of laminopathies, but those mutations had never been linked to ARVC, Kim’s specific heart malfunction. Had she been misdiagnosed? A few months later, Kim stumbled across a new paper by a team of British researchers who had studied 108 people with ARVC and found that four had LMNA mutations (and none of the standard ones). “To the best of our knowledge, this is the first report of ARVC caused by mutations in LMNA,” they wrote. They didn’t know about Kim’s work – they couldn’t have, of course. But she knew. Kim had beaten them to it. “I was so excited, I was running up and down the beach,” she says.

§

When patients get solutions to their own genetic puzzle, it’s always professional geneticists who do the solving. Take James Lupski. He has been studying Charcot–Marie–Tooth for decades, and discovered the first gene linked to the condition. He also has it himself. In 2010 he sequenced his own genome and discovered a previously unidentified mutation responsible for the disease. In other cases anxious parents have been instrumental in uncovering the causes of their kids’ mysterious genetic disorders after long diagnostic odysseys, but only by bringing their cases in front of the right scientists.

Kim, however, was an amateur. And to her, sequencing was not a Hail Mary pass that would – maybe, somehow – offer her answers; it was a way of confirming a carefully researched hypothesis.

“People have been talking about empowering consumers since there was an internet,” says Eric Topol, a geneticist at the Scripps Clinic. “But finally, we’ve reached a point where someone can delve into their condition beyond what the top physicians at the Mayo Clinic could. They couldn’t connect the dots. She did.”

Topol, a self-described “digital medicine aficionado”, argues that Kim is a harbinger of things to come. In his book The Creative Destruction of Medicine, Topol foretells a future where doctors are no longer the gatekeepers of medical information. Advances like personal genetic testing or sensors that measure molecules in the blood will give patients the power to better understand themselves and to exercise more control over their healthcare. Medicine is becoming more democratic.

Kim is a vanguard of that change. She lacked academic knowledge, but she had several advantages over her physicians and other researchers in the field. She had detailed first-hand knowledge of her own symptoms, allowing her to spot connections in the scientific literature that others had missed. She could devote hours to learning everything about her niche disorders – time and focus that no clinician could reasonably spend on a single case. And she had unparalleled motivation: “There’s nothing that engages your curiosity more than being confronted by your death,” she says.

It is also becoming ever easier for that curiosity to lead to discovery. In the past geneticists would try to diagnose patients by looking at their medical history and deciding which genes might be worth sequencing, as Gavrilova tried to do for Kim. The approach makes sense, but it only ever confirms known links between genes and diseases.

One way of finding new links is to sequence a patient’s exome – the 1 per cent of their genome that contains protein-coding genes. It’s cheaper than sequencing a full genome, but allows researchers to hunt for disease-related genes by interrogating every possible suspect simultaneously, without having to whittle down the list first. “Suddenly, we’re finding patients presenting with Disease X who have mutations in genes never previously associated with that disease,” says Daniel MacArthur, a geneticist at Massachusetts General Hospital. “That’s happening in nearly every disease field right now.”

Exome sequencing is now barely more expensive than sequencing much narrower gene panels. MacArthur says that the cost has already fallen below $1,000 and may halve again this year. And once patients have that information, they could use it to find others with the same mutations and check if they have the same symptoms.

Currently, the results from DNA sequencing studies are largely squirrelled away in boutique databases that collate mutations for specific diseases or genes. The ironically named Universal Mutation Database covers mutations in only 34 genes, including LMNA. Broader ones exist, but for decades they have been incomplete, rife with mistakes, or inaccessible, even to other researchers – a sad state of affairs that MacArthur laments as the “single greatest failure in human genetics”. Now, though, the National Institutes of Health are developing an open database called ClinVar that covers all disease mutations. “A lot of us are putting our hopes on this,” says MacArthur. “We need to come up with resources that empower people to make surprising links, which is hard to do if the data are broken up by disease or gene.”

But for every Kim, there are others who research their own conditions and come up with wrong answers. In one study four non-specialist volunteers tried to diagnose 26 cases from the New England Journal of Medicine by Googling the symptoms. They got less than a quarter right. Genetic diseases arguably lend themselves to confusion and misinformation. They are often both debilitating and enigmatic, and getting sequenced can offer little comfort beyond a diagnosis. If mainstream science has no easy answers to offer, many patients will follow any lead, no matter how weak. “There’s a tendency for people to spin very convoluted stories on tenuous threads of evidence. Even scientists do that,” says MacArthur. “I have heard of a lot of rare-disease patients who come up with hypotheses about their disease, and very few turn out to be correct.”

Even Kim’s tale could have taken a different turn. Last year, a team from the Baylor College of Medicine sequenced the exomes of 250 people with suspected genetic disorders, and found that four of them had two diseases caused by mutations in different genes. In other words, Kim’s hunch about her two diseases sharing a common root could well have been wrong. Lightning does occasionally strike twice.

“We almost always have to spend time with patients decoding and recoding the impression that they’ve acquired about their disease from their own homework,” says Ackerman. Kim was an exception, he says, and her other physicians echo that view. She is unique. She is one-of-a-kind. She is extraordinary. High praise, but it conceals the implicit suggestion that she is an outlier and will continue to be.

§

“Bullshit,” says Kim. “I hear this all the time: that I’m an exception. That the patient of the future is not going to do what I did.” She bristles at the very suggestion. “I almost take offence when I hear that what I’ve done is exceptional.”

We are talking over coffee at La Ventana. This is her fifth winter here, and she and CB have just celebrated their 30th wedding anniversary. CB leans back against a wall, quiet and contemplative. Kim sits forward, animated and effusive. She’s drinking decaf because of her heart, but it’s not like she needs the caffeine. “Take Rodney Mullen. He’s a real genius,” she says. Mullen is not a figure from science or medicine. He is, in fact, a legendary skateboarder, famous for inventing mind-blowing tricks that previously seemed impossible. One of them is actually called the ‘impossible’. “He executes these movements that defy reason, films them and publishes them on YouTube,” Kim says. “And inevitably, within a few weeks, someone will send him a clip saying: This kid can do it better than you. He gave that trick everything he had, he’s pulling from all of his experience, and here’s this kid who picks it up in a matter of weeks. Because he learned that it’s possible to do that. Rodney just acts as a conduit. He breaks barriers of disbelief.”

Her protestations aside, Kim is unique. Throughout her life she had built up a constellation of values and impulses – endurance, single-mindedness, self-reliance and opposition to authority – that all clicked in when she was confronted with her twin diagnoses. She was predisposed to win. Not everyone is. But as genetic information becomes cheaper, more accessible and more organised, that barrier may lower. People may not have to be like Kim to do what she did.

Kim isn’t cured. Her LMNA discovery offered her peace of mind but it did not suggest any obvious treatments. Still, she has made a suite of dietary changes, again based on her own research, which she feels have helped to bring her nervous symptoms under control. Some are generic, without much hard science behind them: she eats mostly organic fruit, vegetables, nuts and seeds, and avoids processed food. Others are more tailored. She drinks ginger tea because it thins the blood – she says that many people with laminopathies have problems with clots. Whether her choices are directly slowing the progress of her diseases or triggering a placebo effect, she is fit and happy. Her defibrillator hasn’t shocked her in months. And, of course, she still exercises constantly.

Up the hill from the beach we can see the little yellow house where she wrote the 36-page booklet that put together all her research. It convinced her doctors, yes, but it did even more. She showed it to her brother, now an anaesthesiologist, and it allowed them to reconcile. “It’s like I’ve finally done something worthy with my life,” Kim says. “He told me I’d done some really good research and that I’d missed my calling as a medical researcher. I told him I think I’ve been doing exactly what I needed to do.”

§

Main references

A booklet by the British Heart Foundation that describes arrhythmogenic right ventricular cardiomyopathy (ARVC). 

An overview of Charcot–Marie–Tooth disease by the NHS.

The 2012 paper by the team of British researchers who studied 108 people with ARVC and found that four had LMNA mutations.

The 2008 study in which non-specialist volunteers tried to diagnose 26 cases from the New England Journal of Medicine by Googling the symptoms. [PDF]

Read the full article. This article was commissioned by Mosaic, a digital publication from the Wellcome Trust dedicated to exploring all strands of the science of life. It is reproduced under a Creative Commons Attribution 4.0 International Licence.

Fox via YouTube
Show Hide image

Are smart toys spying on children?

If you thought stepping on a Lego was bad, consider the new ways in which toys can hurt and harm families.

In January 1999, the president of Tiger Electronics, Roger Shiffman, was forced to issue a statement clearing the name of the company’s hottest new toy. “Furby is not a spy,” he announced to the waiting world.

Shiffman was speaking out after America’s National Security Agency (NSA) banned the toy from its premises. The ban was its response to a playground rumour that Furbies could be taught to speak, and therefore could record and repeat human speech. “The NSA did not do their homework,” said Shiffman at the time.

But if America’s security agencies are still in the habit of banning toys that can record, spy, and store private information, then the list of contraband items must be getting exceptionally long. Nearly 18 years after TE were forced to deny Furby’s secret agent credentials, EU and US consumer watchdogs are filing complaints about a number of WiFi and Bluetooth connected interactive toys, also known as smart toys, which have hit the shelves. Equipped with microphones and an internet connection, many have the power to invade both children’s and adults’ private lives.

***

“We wanted a smart toy that could learn and grow with a child,” says JP Benini, the co-founder of the CogniToys “Dino”, an interactive WiFi-enabled plastic dinosaur that can hold conversations with children and answer their questions. Benini and his team won the 2014 Watson Mobile Developer Challenge, allowing them to use the question-answering software IBM Watson to develop the Dino. As such, unlike the “interactive” toys of the Nineties and Noughties, Dino doesn’t simply reiterate a host of pre-recorded stock phrases, but has real, organic conversations. “We grew it from something that was like a Siri for kids to something that was more conversational in nature.”

In order for this to work, Dino has a speaker in one nostril and a microphone in the other, and once a child presses the button on his belly, everything they say is processed by the internet-connected toy. The audio files are turned into statistical data and transcripts, which are then anonymised and encrypted. Most of this data is, in Benini’s words, “tossed out”, but his company, Elemental Path, which owns CogniToys, do store statistical data about a child, which they call “Play Data”. “We keep pieces from the interaction, not the full interaction itself,” he tells me.

“Play Data” are things like a child’s favourite colour or sport, which are used to make a profile of the child. This data is then available for the company to view, use, and pass on to third parties, and for parents to see on a “Parental Panel”. For example, if a child tells Dino their favourite colour is “red”, their mother or father will be able to see this on their app, and Elemental Path will be able to use this information to, Benini says, “make a better toy”.

Currently, the company has no plans to use the data with any external marketers, though it is becoming more and more common for smart toys to store and sell data about how they are played with. “This isn’t meant to be just another monitoring device that's using the information that it gathers to sell it back to its user,” says Benini.

Sometimes, however, Elemental Path does save, store, and use the raw audio files of what a child has said to the toy. “If the Dino is asked a question that it doesn’t know, we take that question and separate it from the actual child that’s asking it and it goes into this giant bucket of unresolved questions and we can analyse that over time,” says Benini. It is worth noting, however, that Amazon reviews of the toy claim it is frequently unable to answer questions, meaning there is potentially an abundance of audio saved, rather than it being an occasional occurrence.

CogniToys have a relatively transparent Privacy Policy on their website, and it is clear that Benini has considered privacy at length. He admits that the company has been back and forth about how much data to store, originally offering parents the opportunity to see full transcripts of what their child had been saying, until many fed back that they found this “creepy”. Dino is not the first smart toy to be criticised in this way.

Hello Barbie is the world’s first interactive Barbie doll, and when it was released by Mattel in 2015, it was met with scorn by parents’ rights groups and privacy campaigners. Like Dino, the doll holds conversations with children and stores data about them which it passes back to the parents, and articles expressing concerns about the toy featured on CNN, the Guardian, and the New York Times. Despite Dino’s similarities, however, Benini’s toy received almost no negative attention, while Hello Barbie won the Campaign for a Commercial-Free Childhood’s prize for worst toy of the year 2015.

“We were lucky with that one,” he says, “Like the whole story of the early bird gets the worm but the second worm doesn’t get eaten. Coming second on all of this allowed us to be prepared to address the privacy concerns in greater depth.”

Nonetheless, Dino is in many ways essentially the same as Hello Barbie. Both toys allow companies and parents to spy on children’s private playtimes, and while the former might seem more troubling, the latter is not without its problems. A feature on the Parental Panel of the Dino also allows parents to see the exact wording of questions children have asked about certain difficult topics, such as sex or bullying. In many ways, this is the modern equivalent of a parent reading their child's diary. 

“Giving parents the opportunity to side-step their basic responsibility of talking to, engaging with, encouraging and reassuring their child is a terrifying glimpse into a society where plastic dinosaurs rule and humans are little more than machines providing the babies for the reptile robots to nurture,” says Renate Samson, the chief executive of privacy campaign group Big Brother Watch. “We are used to technology providing convenience in our lives to the detriment of our privacy, but allowing your child to be taught, consoled and even told to meditate by a WiFi connected talking dinosaur really is a step in the wrong direction.”

***

Toy companies and parents are one thing, however, and to many it might seem trivial for a child’s privacy to be comprised in this way. Yet many smart toys are also vulnerable to hackers, meaning security and privacy are under threat in a much more direct way. Ken Munro, of Pen Test Partners, is an ethical hacker who exposed security flaws in the interactive smart toy “My Friend Cayla” by making her say, among other things, “Calm down or I will kick the shit out of you.”

“We just thought ‘Wow’, the opportunity to get a talking doll to swear was too good,” he says. “It was the kid in me. But there were deeper concerns.”

Munro explains that any device could connect to the doll over Bluetooth, provided it was in range, as the set-up didn’t require a pin or password. He also found issues with the encryption processes used by the company. “You can say anything to a child through the doll because there's no security,” he says. “That means you've got a device that can potentially be used to groom a child and that's really creepy.”

Pen Test Partners tells companies about the flaws they find with their products in a process they call “responsible disclosure”. Most of the time, companies are grateful for the information, and work through ways to fix the problem. Munro feels that Vivid Toy Group, the company behind Cayla, did a “poor job” at fixing the issue. “All they did was put one more step in the process of getting it to swear for us.”

It is one thing for a hacker to speak to a child through a toy and another for them to hear them. Early this year, a hack on baby monitors ignited such concerns. But any toy with speech recognition that is connected to the internet is also vulnerable to being hacked. The data that is stored about how children play with smart toys is also under threat, as Fisher Price found out this year when a security company managed to obtain the names, ages, birthdays, and genders of children who had played with its smart toys. In 2015, VTech also admitted that five million of its customers had their data breached in a hack.

“The idea that your child shares their playtime with a device which could potentially be hacked, leaving your child’s inane or maybe intimate and revealing questions exposed is profoundly worrying,” says Samson. Today, the US Electronic Privacy Information Center (EPIC) said in a statement that smart toys “pose an imminent and immediate threat to the safety and security of children in the United States”. 

Munro says big brands are usually great at tackling these issues, but warns about smaller, cheaper brands who have less to lose than companies like Disney or Fisher Price. “I’m not saying they get it right but if someone does find a problem they’ve got a huge incentive to get it right subsequently,” he says of larger companies. Thankfully, Munro says that he found Dino to be secure. “I would be happy for my kids to play with it,” he says. “We did find a couple of bugs but we had a chat with them and they’re a good bunch. They aren’t perfect but I think they’ve done a hell of a lot of a better job than some other smart toy vendors.”

Benini appears alert to security and the credibility it gives his company. “We took the security very, very seriously,” he says. “We were still building our systems whilst these horror stories were coming about so I already set pipelines and parameters in place. With a lot of devices out there it seems that security takes a backseat to the idea, which is really unfortunate when you’re inviting these devices into your home.”

As well as being wary of smaller brands, Munro advises that parents should look out for Bluetooth toys without a secure pairing process (ie. any device can pair with the toy if near enough), and to think twice about which toys you connect to your WiFi. He also advises to use unique passwords for toys and their corresponding apps.

“You might think ‘It's just a toy, so I can use the same password I put in everything else’ – dog’s name, football club, whatever – but actually if that ever got hacked you’d end up getting all your accounts that use that same password hacked,” he says.

Despite his security advice, Munro describes himself as “on the fence” about internet-connected smart toys as a whole. “Most internet of things devices can be hacked in one way or another,” he says. “I would urge caution.”

***

Is all of this legal? Companies might not be doing enough ethically to protect the privacy of children, but are they acting responsibly within the confines of the law?

Benini explains that Dino complies with the United States Children's Online Privacy Protection Act (COPPA) of which there is no real equivalent in the UK. COPPA says that companies must have parental permission to collect personal information over the internet about children under 13 years of age. “We’ve tried to go above and beyond the original layout of COPPA,” says Benini, when describing CogniToys transparent privacy documents. Parents give their consent for Elemental Path to collect their children’s data when they download the app that pairs with the toy.

Dino bears a striking similarity to Amazon Echo and Google Home, smart speakers that listen out for commands and questions in your home. Everything that is said to Amazon Echo is recorded and sent to the cloud, and an investigation by the Guardian earlier this year discovered that this does not comply with COPPA. We are therefore now in a strange position whereby many internet of things home devices are legally considered a threat to a child’s privacy, whereas toys with the same capabilities are not. This is an issue because many parents may not actually be aware that they are handing over their children’s data when installing a new toy.

As of today, EU consumer rights groups are also launching complaints against certain smart toys, claiming they breach the EU Unfair Contract Terms Directive and the EU Data Protection Directive, as well as potentially the Toy Safety Directive. Though smart toys may be better regulated in Europe, there are no signs that the problem is being tackled in the UK. 

At a time when the UK government are implementing unprecedented measures to survey its citizens on the internet and Jeremy Hunt wants companies to scour teens’ phones for sexts, it seems unlikely that any legislation will be enacted that protects children’s privacy from being violated by toy companies. Indeed, many internet of things companies – including Elemental Path – admit they will hand over your data to government and law enforcement officials when asked.

***

As smart toys develop, the threat they pose to children only becomes greater. The inclusion of sensors and cameras means even more data can be collected about children, and their privacy can and will be compromised in worrying ways.

Companies, hackers, and even parents are denying children their individual right to privacy and private play. “Children need to feel that they can play in their own place,” says Samson. It is worrying to set a precedent where children get used to surveillance early on. All of this is to say nothing of the educational problems of owning a toy that will tell you (rather than teach you) how to spell “space” and figure out “5+8”.

In a 1999 episode of The Simpsons, “Grift of the Magi”, a toy company takes over Springfield Elementary and spies on children in order to create the perfect toy, Funzo. It is designed to destroy all other toys, just in time for Christmas. Many at the time criticised the plot for being absurd. Like the show's prediction of President Trump, however, it seems that we are living in a world where satire slowly becomes reality.

Amelia Tait is a technology and digital culture writer at the New Statesman.