When she first began talking to her peers in the House of Lords about the rights of children on the internet, Baroness Kidron says she looked like “a naysayer”, like someone who was “trying to talk about wooden toys” or, in her husband’s words, like “one middle-aged woman against Silicon Valley”. It was 2012 and the film-maker and recently appointed life peer was working on her documentary InRealLife, spending “hundreds of hours in the bedrooms of children” to discover how the internet affects young lives. What she saw disturbed her.
“I did what they were doing – gaming, falling in love, watching pornography, going to meet-ups, making music – you name it, it happened,” Beeban Kidron says. The film explored everything from children’s exposure to porn, to rampant online bullying, to the way privacy is compromised online. But Kidron noticed that one thing underpinned it all: on the internet, nobody knows you’re a kid. “Digital services and products were treating them as if they were equal,” she says. “The outcome of treating everyone equally is you treat a kid like an adult.”
Almost a decade later, Kidron has pushed through a Children’s Code that hopes to change this landscape for ever. The Age Appropriate Design Code, an amendment to the 2018 Data Protection Act, came into effect this month. It requires online services to “put the best interests of the child first” when designing apps, games, websites and internet-connected toys that are “likely” to be used by kids.
In total, there are 15 standards that companies need to adhere to in order to avoid being fined up to 4% of their global turnover. These include offering “bite-size” terms and conditions for children; giving them “high privacy” by default; turning off geolocation and profiling; and avoiding “nudge techniques” that encourage children to turn off privacy settings. The code, which will be enforced by the Information Commissioner’s Office (ICO), also advises against “using personal data in a way that incentivises children to stay engaged”, such as feeding children a long string of auto-playing videos one after the other.
The code was introduced in September 2020, but offered companies a 12-month transition period, in this time the world’s tech giants have seemingly begun responding to the sting of Kidron’s sling. Instagram now prevents adults from messaging children who don’t follow them on the app, while anyone under 16 who creates an account will have it set to private by default. TikTok has implemented a bedtime for notifications; teens aged 13-15 will no longer be pinged after 9pm. Meanwhile, YouTube has turned off autoplay for users aged 13-17, while Google has blocked the targeted advertising of under-18s.
But hang on, why does TikTok’s bedtime only apply to those 13 and over? Are toddlers OK to use the app until 2am? You’ve just spotted the first flaw in the plan. While social media sites require users to be at least 13 to sign up for their services (in line with America’s 21-year-old Children’s Online Privacy Protection Act), a quick glance at reality shows that kids lie about their age in order to snap, share and status-update. Creating a system in which children can’t lie, by, for example, necessitating that they provide ID to access an online service, ironically risks compromising their privacy further.
“There is nothing that stops us having a very sophisticated age-check mechanism in which you don’t even know the identity of the person, you just know that that they’re 12,” Kidron argues, pointing to a report on age verification that she recently worked on with her organisation 5Rights Foundation, entitled But how do they know it is a child?. Third-party providers, for example, could confirm someone’s identity without passing on the data to tech giants, or capacity testing could allow websites to estimate someone’s age based on whether they can solve a puzzle (no prizes for figuring out the numerous ways that could go wrong).
Whatever the solution, Kidron is currently working on a private member’s bill that sets minimum standards of age assurance, thereby preventing companies from choosing their own “intrusive, heavy handed or just terrible, lousy, and ineffective” techniques.
How did Kidron go from looking like a “naysayer” to changing the landscape so drastically? Kidron began making documentaries in the 80s before working in Hollywood (most notably directing the Bridget Jones sequel The Edge of Reason). After becoming a baroness, she founded the 5Rights Foundation to fight for children’s digital rights. She says she had her “early adopters” in parliament, including the archbishop of York, Stephen Cottrell, Conservative peer Dido Harding and Liberal Democrat peer Timothy Clement-Jones. “That was my gang,” Kidron says, but others remained sceptical for years. “The final set of people only came on board this summer, once they saw what the tech companies were doing.”
The Children’s Code as a whole defines a child as anyone under 18, in line with the United Nations Convention on the Rights of the Child (UNCRC). For Kidron, it’s about much more than privacy – “a child’s right to unfettered access to different points of view is actually taken away by an algorithmic push for a particular point of view,” she argues, also noting that the right to the best possible health is removed when companies store and sell data about children’s mental health. “It’s nothing short of a generational injustice,” she says. “Here was this technology that was purporting to be progressive, but in relation to children it was regressive – it was taking away the existing rights and protections.”
How did these claims go down in Silicon Valley? Conversations with executives were surprisingly “very good and productive”, according to Kidron, but she ultimately realised that change would have to be forced upon tech companies. “They have an awful lot of money to have an awful lot of very clever people say an awful lot of things in an awful lot of spaces. And then nothing happens,” she says. “Anyone who thinks that the talk itself is going to make the change is simply wrong.”
And yet while companies must now comply with the code, even Kidron admits, “they have to comply in ways that they determine”. TikTok’s bedtime, for example, seems both arbitrary and easy to get around (children are well versed in changing the date and time on their devices to proceed in video games). Yet Kidron says the exact o’clock is irrelevant – the policy is about targeting sleeplessness in children, which in turn enables them to succeed at school. “These things seem tiny… but they’re not. They’re about the culture and they’re about how children live.”
As for children working their way around barriers, Kidron notes that transgression is part of childhood, but “you have to allow kids to transgress, you can’t just tell them it’s really normal”. “The problem we have is kids who are eight are looking at hardcore, violent, misogynistic porn and there’s no friction in the system to say, ‘Actually, that’s not yours.’”
Yet problems also arise when we allow tech companies, not parents, to set boundaries for our children. In 2017, YouTube came under fire after its parental controls blocked children from seeing content made by LGBTQ+ creators (YouTube initially apologised for the “confusion” and said only videos that “discuss more sensitive issues” would be restricted in the future). Kidron says she’s “not a big takedown freak” and is “committed to the idea that children have rights to participate”, but can the same be said of companies hoping to avoid fines? Numerous American websites remain inaccessible in Europe after the implementation of General Data Protection Regulation (GDPR) laws in 2018, with companies preferring to restrict access rather than adapt.
For now, it remains to be seen how the Children’s Code will be enforced in practice; Kidron says it’s “the biggest redesign of tech since GDPR”, but in December 2020 a freedom of information request revealed that more than half of GDPR fines issued by the ICO remain unpaid.
Still, Kidron is certain of one thing: that tech companies are “disordering the world” with their algorithms – “making differences of their terms for people who are popular and have a lot of followers versus those who are not” and “labelling things that get attention without really thinking about what that attention is about”. These are prescient remarks: a day after we speak, the Wall Street Journal revealed that Facebook has a program that exempts high-profile users from its rules and has also published internal studies demonstrating that Instagram is harmful to teens. One internal presentation slide read: “We make body image issues worse for one in three teen girls.” Instagram’s head of public policy responded to the report in a blog post, writing: “The story focuses on a limited set of findings and casts them in a negative light.”
Whether or not Kidron was once “one middle-aged woman against Silicon Valley”, today she has global support. The recent changes implemented by social media companies are not just UK-based, but have been rolled out worldwide. Kidron says her code is a Trojan horse, “starting the conversation that says, you can regulate this environment”.
But this Trojan horse is only beginning to open up. “We had 14 Factory Acts in the 19th century on child labour alone,” Kidron says, adding that the code is likely to be the first of many more regulations to come. “I think today we air punch,” she says, when asked how it feels to have led the charge for change. “Tomorrow, we go back to work.”
The price of a 2GB Raspberry Pi 4 single-board computer is going up $10, and its supply is expected to be capped at seven million devices this year due to the ongoing global chip shortage.
Demand for components is outstripping manufacturing capacity at the moment; pre-pandemic, assembly lines were being red-lined as cloud giants and others snapped up parts fresh out of the fabs, and the COVID-19 coronavirus outbreak really threw a spanner in the works, so to speak, exacerbating the situation.
Everything from cars to smartphones have felt the effects of supply constraints, and Raspberry Pis, too, it appears. Stock is especially tight for the Raspberry Pi Zero and the 2GB Raspberry Pi 4 models, we’re told. As the semiconductor crunch shows no signs of letting up, the Raspberry Pi project is going to bump up the price for one particular model.
The 2GB Raspberry Pi 4 will now once again set you back $45, an increase of $10 from its previous retail price. It used to be $45, then was brought down to $35 early last year when the 1GB model was discontinued. Now it’s back up again. This is the first time the project has hiked its prices, the trading arm of the Raspberry Pi Foundation said.
Don’t worry, however, the bump is said to be temporary and the module will eventually return to its original price of $35, company CEO Eben Upton announced on Wednesday.
The 4GB Raspberry Pi 4 and 8GB Raspberry Pi 4 versions will remain at $55 and $75, respectively. For those relying on a supply of $35 2GB boards, the project will bring back those 1GB Raspberry Pi 4 modules, priced $35.
“This provides a degree of choice: less memory at the same price; or the same memory at a higher price,” said Upton. 2GB for $45 or 1GB for $35. A choice, but not one people might expect.
“As many of you know,” he continued, “global supply chains are in a state of flux as we (hopefully) emerge from the shadow of the COVID-19 pandemic. In our own industry, semiconductors are in high demand, and in short supply: the upsurge of demand for electronic products for home working and entertainment during the pandemic has descended into panic buying, as companies try to secure the components that they need to build their products … At Raspberry Pi, we are not immune to this.”
The project is expected to make around seven million of its computer boards total this year, maintaining the same level of production as last year as the pandemic took hold of the world. This is unlikely to increase much next year either, Upton said. Judging from his explanation, this figure is lower than hoped: “Despite significantly increased demand, we’ll only end up making around seven million units in 2021.”
Pis containing 40nm chips will feel the chip crunch the hardest over the next year, meaning there will be limited supplies of devices older than the current generation of Raspberry Pi 4, Raspberry Pi 400, or Compute Module 4.
“In allocating our limited stocks of 40nm silicon, we will prioritise Compute Module 3, Compute Module 3+, and Raspberry Pi 3B, and deprioritise Raspberry Pi 3B+ … Our guidance to industrial and embedded users of Raspberry Pi 3B+ who wish to optimise availability in 2022 is to begin migrating your designs to the 1GB variant of Raspberry Pi 4,” Upton said.
The biz expects to be able to make enough systems using 28nm silicon – namely the Raspberry Pi 4 and Compute Module 4 – over the next 12 months to hold their price… bar the aforementioned 2GB model.
“These changes in pricing are not here to stay. As global supply chain issues moderate, we’ll keep revisiting this issue, and we want to get pricing back to where it was as fast as we can,” Upton concluded. ®
UK headquartered Swoop was one of three finance companies to have received funding from RBS, which has previously given the start-up £5m in 2019.
Irish start-up Swoop Finance has received £2.5m from a fund established by banking giant RBS.
In 2019, it was awarded £5m by the banking firm, which accepted a £45bn bailout from the UK government at the height of the financial crisis in 2018. The bailout programme came with the condition that RBS would set up a £775m fund to boost competition in the region’s finance sector.
Swoop is one of three companies to have benefitted from that fund, with the others being UK finance companies Codat and Cashplus. The three start-ups will receive a combined £12.5m in grants from RBS.
Codat and Cashplus will both receive £5m from the fund.
Swoop was founded in 2017 by former KPMG chartered accountant and corporate financier Andrea Reynolds along with Ciarán Burke. Reynolds spoke at Silicon Republic’s Future Human event last year about the process of launching Swoop. She said she founded it after she spotted a gap in the market for a virtual “finance buddy” aimed at SMEs seeking financial advisers and lenders.
Today, Swoop is headquartered in the UK and it employs around 60 people. It recently launched in Canada, adding to its existing locations in Dublin, London and Sydney.
The fintech’s backers include Enterprise Ireland and Velocity. It has raised around €1.6m so far. Speaking last year, Reynolds said the pandemic’s digitisation of the finance industry – and most other industries – had benefitted the company.
She added that the ongoing changes in the industry would hopefully “democratise finance” and “open up opportunities” to companies seeking funding no matter where they are located.
“The future is that you won’t need to know who the lender is,” Reynolds said.
“All decisions will be made through your data and you’ll get those decisions instantly. So you could have a lender in Barcelona lending to a business in Ballyjamesduff, for example. It won’t matter where you are. It’s what your profile is and does it match to their algorithm.
“This means it’ll open up opportunities. It’ll democratise finance further because businesses, regardless of where they’re located, will not be disadvantaged. Everybody will have this at their fingertips,” she added.
Reynolds said she had seen “a 30pc increase in businesses moving online” during the Covid-19 pandemic.
I’m going to try to convince you that you should care about exactly how many people watched the moral panic-inducing hit Netflix series Squid Game. Yes, I know there’s a lot going on in the world. But bear with me: I think this really matters regarding how we understand our culture – and the balance of power in a media business where data is king.
(As a treat, if you stick with this newsletter then further down I’ll tell you about some of the biggest flops that Netflix would prefer you didn’t know about.)
In the past it was simple to find out how many people watched a popular television show. Audience figures for traditional television broadcasts have been produced in a similar way for decades. Research companies recruit a group of households considered to be statistically representative of the general population (in the UK this is done by Barb, in the US by Nielsen) and then their viewing habits are monitored, often using a box attached to their television set.
This data is then processed and used to produce industry-standard television ratings that can make or break careers. Journalists love these figures because you can make narratives out of them! It’s why you see headlines in news outlets about how half of the UK watched a football match, or how no one watched a new rightwing news channel.
These figures are made public, in part, because commercial television channels have advertisers. And advertisers need to know their adverts are actually being watched, so they need reliable and trustworthy numbers produced by a third-party organisation. Sure, this survey system is flawed, but broadly speaking it is equally flawed for everyone. You can tell if a programme on BBC is much more popular than a show on ITV, and you can tell if a particular drama massively outperformed what you’d expect.
Then Netflix and subscription streaming services came along. They don’t have advertisers. Their aim is to hook, retain, and encourage customers to keep using their service until it becomes so ingrained in their lives that they can never stop paying their monthly subscription fee. Core to working out how to do this is the data they collect on you.
Because Netflix knows exactly what shows you watch. They know how many seconds you lasted with each programme, when you got bored, what you put on instead when you got bored, and exactly what time of night you were watching that smutty foreign series. And it’s really not in Netflix’s interest to share this information with journalists, their rivals, or with the people who make the shows.
Which brings us back to the original question: How many people watched Squid Game? And why does it matter?
Well, if you believe Netflix, who occasionally drip-feed out positive ratings stories when it suits them, by last night Squid Game had been watched by 142 million households, making it one of the biggest hits ever.
But we’ve only got Netflix’s word to go on for that figure. And even then, Netflix currently defines a viewer as someone who watched the first two minutes of a show’s opening episode. Did you put Squid Game on for a few minutes to check out the hype then get bored? Well, you might be surprised to find you’re counted to be just as much a “fan” of the show as someone who watched all nine episodes back-to-back.
Journalistically, it’s a challenge. We end up having to accept Netflix’s word for the figures they provide because there’s simply no other option. It also enables the streaming outlets to selectively publish the narrative that they want to construct. It’s sexy and cool to trumpet your investment in high-end original drama. (And hell, Netflix really is investing incredible sums in high-end original drama!) It’s less sexy to admit that your critically acclaimed show was a ratings flop and people just want to watch endless repeats of Grand Designs.
What’s more, it warps our perceptions of audiences and what is popular in culture. Is a Netflix drama more popular than a BBC drama? Possibly. This may have enormous implications for the future of whether we still need the licence fee. Does the public really engage with Oscar-nominated state-of-the-nation films or secretly sit there watching another Adam Sandler release? With the culture wars grinding on, it’s probably worth knowing. What are the truly unifying television moments that bind a society together? It’s hard to be sure. Because we can’t get the data out of Netflix.
The truth is out there
Except … one small family business based in Bristol has worked out how to do just that. The staff at Digital i, an analytics firm, realised that while Netflix won’t release viewing figures, it does release data to members of the public about their personal viewing history.
(It’s true, you can see an overview of your recent Netflix viewing history, or you can download every bit of data that Netflix holds on you by visiting this link. In my case, it reveals that I was really binge-watching an awful lot of episodes of The Good Wife in 2015.)
Digital i realised that if they could convince thousands individuals to willingly hand over this personal viewing history in return for a small payment, the company can effectively create a statistically rigorous survey panel, then use this to create audience “ratings” for Netflix shows and sell this data to rivals. At the moment they have users signed up in five major European countries but they hope to expand globally.
“We’re trying to level the playing field for Netflix competitors,” said Sophia Vahdati from the company, who says their customers include the likes of BBC and ITV.
Her company has shone a light on one of Netflix’s biggest secret: how much of their audience is viewing endless repeats of old shows, because people binge high-profile original series in such a short period of time.
“The biggest thing that isn’t mentioned in the hype is how important sitcoms are to retaining Netflix subscribers,” she said, highlighting the availability of Brooklyn 99 and Big Bang Theory as just as core to Netflix’s offering as their buzzy acclaimed shows.
Here’s some of the findings of their Digital i’s data from its UK audience research that she shared with the Guardian:
British Netflix users spent more time watching old episodes of Friends in 2020 than watching big-budget original series the Crown.
The three most popular new releases in the UK during August were Clickbait (watched by 2.34m Netflix accounts), Hit & Run (2.1m households), and The Chair (1.64m). These are high ratings but Channel 5 can top them.
Sex Education Series 3 was released on the same day as Squid Game and performed just as well in Europe – but has had a fraction of the hype.
Shows such Bridgerton, Afterlife and The Queen’s Gambit were all hitting over 80% completion rates in the UK – meaning people were hooked and watched to the end of each series.
At the other end of the market, the five shows with the worst series completion rates were The Dark Crystal: Age of Resistance (just 35% of viewers finished it), What/If (45%), The Irregulars (53%), White Lines (56%), and Sex/Life (56%) – which explains why most of them were cancelled.
Any film that is watched to the end by 70% of people is a success. Martin Scorsese’s big-budget much-hyped Irishman? That struggled, on their metrics.
People now watch original series in a very short space of time – about a quarter of people who watched Squid Game finished it within two days.
Even though Netflix and Amazon Prime Video are not far apart in terms of signed-up users, Netflix dwarfs Amazon when it comes to people actually watching their content.
Oh and almost no one chooses to watch the credits nowadays. Sorry to everyone who made the programmes, we’ve already autoplayed the next episode.
So why does all this matter?
A lack of transparency changes the balance of of power when it comes to small companies negotiating with a global giant such as Netflix.
One independent producer who sold a film to Netflix suspects their release performed well, based on online reaction. But they told me that they just don’t know: “Netflix doesn’t usually give producers information about viewing figures of films they made – which is both frustrating and very disempowering for producers trying to negotiate funding for the next one, with them or anyone else.”
And for Squid Game? Digital i reckons 79% of Europeans with Netflix on their research panel watched at least one episode within the first fortnight of its release – with half making it all the way to the end in that time. So it really is a massive hit. Just perhaps not quite as big as Netflix’s own figures would suggest.
Last night, the streaming company announced that they would slightly change the metrics they use and drop the “two minutes watched” measure in favour of total hours watched. But it’s still in the company’s gift when they make the information public.
Vahdati says her company’s data shows how the streamer can selectively release data to shape the narrative about their output: “The originals are punchy, sharp and aesthetically innovative. But at the heart of it we haven’t become a nation who like to be challenged all the time with foreign-language dramas.”
Oh – and if you’re one of the many Squid Game viewers, then no spoilers please. I’m still only two episodes in.
If you want to read the complete version of this newsletter please subscribe to receive TechScape in your inbox every Wednesday.