19 – Not Just A Tech Problem

September 24, 2020 in Weekly Newsletter

This week we were reminded that many of the problems we’re seeing emerge with tech are just new iterations of problems we’ve always faced.

Alexis Magrigal in the Atlantic looked a series of killings widely reported as “caused by WhatsApp”.

“This year has been presented as an epidemic of violence, aided and abetted, even caused, by WhatsApp. The narrative slotted neatly into the broader discussion of Big Tech’s failures, the corrosiveness of social media, and the crises of misinformation across the world. After all, WhatsApp usage has exploded in India over the past few years, across city and country, rich and poor. Two hundred million Indians now use WhatsApp. Communal violence has been on the rise, going from 751 incidents resulting in 97 deaths in 2015 to 822 incidents and 111 deaths in 2017”

He makes a compelling case that many are just bad humans doing what they’ve always done, just now on WhatsApp.

Another paper published in Science points to the fact that “fake news” constitutes just 0.14% of an average American’s daily media diet.

Closer to home we had several reminders that disinformation and sensationalism is a wider societal problem, with 2FM looking for a debate on the efficacy of masks, and RedFM broadcasting advice that Vitamin C can cure Covid.

I don’t highlight these to encourage us to treat tech related (or tech enhanced) problems any less seriously or urgently, but just to remind that a narrow focus on just the tech involved can blind us to the deeper root cause, making our diagnoses and ultimately our solutions less effective.

📰 News

Are Twitter’s Images Biased? People can upload images of any shape or size to Twitter, but Twitter crops a small section of the image to use as a preview. In 2018 they introduced “smart auto-cropping“, a machine learning algorithm that tries to use the most interesting part of the image in the preview. Several users noticed this week that, in images with black people and white people, the algorithm seems to always focus on the white person. EG 1. EG 2.

Twitter responded fairly quickly, saying that they did test for bias when they launched, but they’ll investigate again. Another person did a test with a random sample and couldn’t recreate the effect, so it may be a case that the only examples becoming popular are the ones that seem to show bias, rather than the ones that don’t.

The whole incident is a great example for why we need more frameworks for auditing the results of automated processes. Their effect matters far more than their intent. In this case Twitter’s engineers just wanted to know “what’s interesting?” in the image, but some checks for “Is this automating racism?” “Is this automating sexism?” would still have been useful. Link.

Co-working. Apparently there’s no 400 remote working spaces in Ireland, mostly in rural areas. Rachel Lavin profiled a few of them in the Business Post. Link.

Brexit. The Johnson Government are detailing their post-Brexit plans to pick the future winners in emerging technologies, give them state-aid and create a world-leading British tech sector. Besides the difficulty this will cause in reaching a trade agreement with the EU, and the fact that all of us would love the clairvoyance necessary to pick the next Google, there’s some merit in direct funding of new technologies with public benefit. The US tech sector benefits from a lot of R&D done as publicly funded military funding, so they could achieve similar effects here, if done through the NHS for example. Then again, you probably don’t need to leave the EU to do that. Link.

💡 Interesting Links

Misinformation. An interview with Bill Gates about misinformation online. Over 20 mins he highlights many of the areas where we know the problems, but don’t yet have many concrete solutions. Link [YouTube].

Fixing bias. Algorithms can help us uncover and address systemic bias. Here’s one good example from France, where the allocation of daycare places is in the gift of local politicians, who are fighting against the algorithm that will do this more fairly. Link.

Oldspiracy. Forget 5G, in 1890 the New York Herald reported a theory that Russian Influenza was caused by the newfangled electricity. Link.

The first death sentence over zoom. As if 2020 couldn’t get more grim. Link.

Buying myself back” Model Emily Ratajkowski talks about what it means to own your own image. An excellent read. Link.

Newsletter 18 – The Fracturing Internet

September 18, 2020 in Weekly Newsletter

Image via NYT.

Two big international stories this week might seem quite disparate, but have the same underlying cause. It looks like Chinese owned TikTok will be bought by Oracle, to placate the Trump administration. Nearer to home the Irish Data Protection Commissioner has ordered Facebook to stop sending private data from the EU to the US, in a move that might break much of how the internet currently works.

Both are indicators that we have moved out of the “beginning” stage of the internet, where every country’s goal was just to get people and businesses “online”. Now we are embarking on the stage where, given that everyone is connected, governments try mould the global internet to match their local cultural norms and legal frameworks.

Because the Us Government can snoop on any US server it wishes, Europe is worried that Facebook and others will be a vector for US security policy, to the detriment of EU citizens. The US is worried that TikTok will be a vector for Chinese foreign policy, to the detriment of US national security.

All of these moves tread a difficult line, trying to impose your worldview to shape a single global internet, without fracturing the internet into different regional versions. China went first many years ago, intentionally fracturing the internet in their region with their “Great Firewall.” Other countries are ramping up their efforts too. India has banned many Chinese companies over the last few months. Australia has released a draft digital news bill that includes mandating 28 day notice before any algorithm changes (the kind that are often deployed on a daily basis).

This has been an emerging trend over the last few years, but in 2020 it has really accelerated. I think it will be a dominant trend for the 2020s, where we struggle with the tension between keeping the benefits of an open, global internet and the desire to make it conform to the norms and rules in our own different parts of the world.

Stories: Oracle likely to buy TikTokFacebook Ireland vs DPCAustralia’s media regulation billIndia bans 118 chines apps.

📰 News

An Post has announced it will be launching its own “Revolut-like” app in October. I saw a lot of Irish techies laugh about this on Twitter, at the notion that the Irish postal service could beat a digital-native like Revolut, but An Post is a fairly serious personal finance organisation. They have Credit Cards, Loans, insurance products and a mobile MVNO. In all of these they’ve been quite smart about how they partner with technical experts to build the product and leverage their own branding and network, so I wouldn’t write them off just yet. Link.

Journalistic privilege. An interesting case – a local journalist filmed some footage of a house on fire. Gardaí suspect fowl play and want to see the contents off his phone. He refused, invoking journalistic privilege, but a judge ruled that there’s no such thing. Link.

The Dublin Inquirer has an excellent piece on Mobility-as-a-Service (MaaS) which several groups are pushing for in Dublin. Once you strip back the technical language, they’re mostly talking about a Leap Card with a monthly subscription, which covers public and private transport, including the likes of Free Now taxis and bleeper bikes. The technical challenge of that sounds fun and interesting, but I’m skeptical about the public benefit of deeply integrating private transport options into a system like this. Link.

TikTok recently switched their legal European HQ to Dublin, which means we have the burden and cost of regulating them, so thankfully we might also be getting some jobs along with that too. The Irish Times reports that they’re looking for property here that could hold up to 5,000 workers. Google employs 8,000 after 15 years here, so take TikTok’s number with a pinch of salt, but it does add weight to the IDA’s claim that getting a company to locate their Data Centre here anchors them, then frequently leads to job creation. I wonder where they’ll all live? Link.

🤔 Interesting Links

Facebook often talk about their platform as a mirror for reality, but Shida Ovide illustrates the many ways that social media platforms shape reality, instead of passively reflecting it. Link.

Rethinking Rocket Internet. I always thought of Rocket Internet, the German startup builder, as a company that just copies others. And it is, but maybe that’s a good thing? It doesn’t have the romantic flair of some startup stories, and as a business owner they scare me, but it’s hard to argue that they’re a net-negative for creating competitive markets. Link.

“Another experiment showing how influential Wikipedia is on the real world: Adding two paragraphs of text & nice pictures to randomly selected articles about small European cities led to an over 9% increase in hotel stays; the edit is worth $190k per year!” Link.

Newsletter 17 – Dublin’s Fair(ly empty) City

September 10, 2020 in Weekly Newsletter

Was 2019 the peak of city-centre office working? Or will 2020 be remembered as a temporary blip when we all go back to the office after Covid?

6 months into the pandemic Google’s decision to abandon plans for a new 2,000 person office in Dublin feels like a strong signal that this may a trend that continues after the pandemic ends. This mirrors similar developments in cities all over the world, with the most notable one being Pinterest’s decision to cancel their lease and pay an $89.5m fine to do so.

“As we analyze how our workplace will change in a post-COVID world, we are specifically rethinking where future employees could be based,” their CFO said. At $89m, it’s a serious bet against the future of the city centre office.

If companies like Google and Pinterest are right, that poses two big questions – Where do people chose to live, when they can work anywhere? And what happens to cities if the office workers don’t return?

On the first question, for now at least, many office workers are getting out of the city. If you can work anywhere, you might as well work somewhere beautiful.

Anecdotally, I’ve heard that the co-working spaces in places like Skibbereen and Tralee are fully booked out these days. Airbnb have had such a strong resurgence over the summer that revenues are now 75% higher than they were at this time in 2019. Hotels are still way down.

It’s hard to know if this trend will hold or if it’s just a fun thing to do during an odd year. Cities are still where people do much more than work – where they socialise, meet romantic partners, have a wider variety of entertainment and food, etc.

Which leads to the second question… what do cities look like if office workers don’t return? In one of the best pieces I read this week, Paul Kearns writes about Dublin that “A capital that prioritises suburban shoppers, commuters and international tourists is now reaping the rewards of its anti-urban living policies.” The pandemic has revealed just how little we’ve invested in Dublin as a city to live in, rather than just one we commute and travel to. Other cities are worrying about the long term emptiness of their business districts, but we might have to worry about the whole city centre.

There’s also the knock on effects. Steve Levine points to the trillion dollar “office economy” that is being killed by remote work. Nobody is queueing in Starbucks. Business travel made up 60% of all airline revenue and (fingers crossed) may not ever be back to those levels (although I’m not sure about this one). I got a sad goodbye email from the barista in our office building last week and I can’t imagine the cleaning staff have been too busy over the last few months.

If the trend holds, those who “feed, transport, clothe, entertain, and shelter people when they are not in their own homes” — will risk losing their jobs, according to Elisabeth Reynolds at MIT. Another indicator that the pandemic is affecting those on lowest incomes the hardest.

I’m not sure that we should spring into action with any long-term plans while the ultimate outcome is still so uncertain, but it certainly gives us opportunity to re-imagine what we should want from our cities, and indeed our suburbs and towns too.

National Digital IDs.

The Economist has a pair of interesting articles this week. In the first they look at how the pandemic is spurring the digitisation of Government. Lots of interesting case studies of public services shifting dramatically online – from wedding registrations, unemployment claims and doctor’s appointments. We have some stellar examples in Ireland – Revenue Online is mostly excellent (just don’t lose your cert), and the Passport application app is a delight.

They also argue, because of this, that Covid-19 strengthens the case for digital ID cards. I’d broadly agree, with all of the caveats they list, including strict legal protections of the data, giving citizens ownership and control of their data and investing heavily in security.

Two of the best practices they recommend – 1) be transparent about it (i.e. call it national id, don’t try sneak it in as something else) and 2) don’t make it mandatory – are of course the exact opposite of what we did here, claiming the Public Services Card wasn’t a national id, then making it mandatory.

Hopefully we learn from the success of initiatives like the Covid Tracker app, which prioritised our data privacy, promoted the benefits of use and invested in building trust, and got very high levels of voluntary adoption as a result.


Mobile data is getting cheap. When I worked in Vodafone, I managed the launch of Prepay mobile broadband in 2008. If memory serves, we made a massive price drop from €5 per mb to €0.5 per mb. This week 48 (Three’s youth brand) launched 100GB (and free calls and text) for €7.99 per month. That’s €0.0000799 per mb! Link.

In news that will make a small number of people very happy, the BlackBerry is back! Coming next year. Link.

The first bit of science on whether or not Contact Tracking apps actually help reduce the spread of Covid has been published by Oxford, which seems to suggest that they have a positive impact at any level of uptake, not just when most people have it installed. Link.

Interesting Links

Abeba Birhane, a cognitive science PhD student at UCD, has published a paper on the “Algorithmic Colonization of Africa”. Link.

On a topic close to my heart – the battle for the future of young men online. “‘Queer Eye’, Jordan Peterson and the battle for depressed men.” Link

A good series of Tweets from Stephen McIntyre at Irish VC Frontline, about why the Google office announcement doesn’t mean Google are pulling out of Ireland. He makes two interesting points: 1) Engineering and Sales are the important jobs at tech companies, the ones with the biggest salaries and the ones least likely to get automated or outsourced. We have lots of them here. 2) We’ve always done well at attracting Inside Sales teams here – the ones that make sales over the phone and close deals with an email. We’ve done less well on the sales teams that do in-person pitches and end deals with a handshake. They go to London. But that’s an opportunity for us, now that business travel and hand-shakes no longer exist. Link.

Election interference guns for hire. A US PR firm was running large scale, coordinated inauthentic behaviour online to influence the elections in Venezuela and Bolivia. Link.

“Should Google’s Ad Market Be Regulated Like the Stock Market?” Link.

Future crimes. A county sheriff in the US implemented a “futuristic program to stop crime before it happens.” Of course, it just lead to innocent people being monitored and harassed. Bad police can just be bad at scale now. Link.

Peter’s Newsletter 16 – Leaving Cert Predictive Grades

September 3, 2020 in Weekly Newsletter

I send pieces like this in a newsletter every Friday morning. Subscribe for free here.

I’m on my holidays in lovely Lahinch this week, so this will be shorter than usual.

Leaving Cert Calculated Grades – A New Natural Experiment

This week it was widely reported that the Leaving Cert predictive grading model has been changed. The minister announced on Tuesday they have “removed the use of school-by-school historical data in the standardisation model.” The department issued their own briefing note suggesting it hasn’t been removed entirely, but dramatically down-weighted.

This is quite an interesting natural experiment we’re running at a country-wide level. For just a single year, to quote the minister “Your school will not determine the results that you get through standardisation this year.” Your school has always determined this, of course, and it will again in the future, but far less so in 2020. The universities will have a higher percentage of students who have had less access to grinds, or smaller class sizes, or after school study, but one would assume have no less potential than their better-off peers.

Early indications from the department are that two trends are emerging:

  1. Overall grades are higher. Teachers have been kinder to students in 2020 than the marking system generally is. This will be tough on people who sat the Leaving in 2019 and are re-applying to the CAO now, competing with an inflated year. Similar for those who choose to sit an exam in November.
  2. The attainment gap between disadvantaged schools and the rest has narrowed this year, which is positive. Being a top performer in a disadvantaged school now means you have a chance of being among the top achievers in the country, which was not the case before.

Calculated grades are released on Monday, and the model used should be made public shortly after that too. Link.

Deliveroo tragedy. In last week’s issue I was talking about the precarious nature of many gig economy workers. On Monday night, tragically, Thiago Osorio Cortes, a 28 year-old deliveroo rider was killed in a hit-and-run in Dublin city. Crowds of Deliveroo workers held a vigil at the spire on Wednesday. Watch. Read more.

How the British government rules by algorithm. The Economist have an interesting piece on how “the over-enthusiastic application of scientific management is weakening the public sector.” I love some data-driven decision making as much as the next guy, but it’s good to see push back on the limitations and negative effects of an over reliance on data-centric approaches. “Targets create three common problems. They produce perverse results when people focus excessively on them. They tempt managers to manipulate numbers. The obsession with measurement diverts people from useful activity to filling in forms.” Read more.

A report on Micromobility (think e-Bikes and e-Scooters) shows that the industry has been hit dramatically by Covid, but should recover entirely and then grow. This is specifically the industry of shared Micromobility – like Dublin Bikes, but electric – which are predicted to be preferable to mass transit, when the cities start to re-populate. Link.

Facebook’s Election Ads. In Ireland there is a moratorium on talking about the election in the press on the day before and of an election. There’s no such moratorium online, or of online ads. Many wondered if Facebook might introduce a similar moratorium for the upcoming US elections. Yesterday they announced a half step in this direction – they won’t allow new ads in the 72 hours before election day. This means that every message that is out there and being promoted will be floating around for a few days before voting day, so in theory can be fact checked, vetted, disputed etc. and try prevent any last minute tom foolery. 72 hours seems like a very short time window for all that to happen. It took months to unravel some of the things that happened in the Brexit campaign, for example. Link.

Peter’s Newsletter 15 – Ireland’s Gig Economies

September 3, 2020 in Weekly Newsletter

I send this update in a newsletter every Friday morning. Subscribe for free here.

In this week’s issue: Gig economy workers, Yelp data predicts covid hotspots, fake guardian headlines, what Europe can teach silicon valley, the UK’s grading algorithm, the human cost of conspiracy theories and most of Scots wikipedia is made up. 

📰 Ideas

How do we classify workers in the gig economy? When I lived in San Francisco one of my housemates used to do “Task Rabbit” jobs the odd evening or weekend for a few extra dollars. He’d usually get a job to assemble some furniture, or move a couch or lift something heavy (he’s a tall guy!). He wasn’t a full time employee, he didn’t have to commit to a certain number of hours. He’d just mark himself available on the app when he wanted some extra spending money. It was the perfect arrangement for him.

But for many others it wasn’t so simple. Taxi drivers who were now Uber or Lyft drivers didn’t always have that luxury, many did it as a full time job, not just a nixer. And many of the platforms demanded more in return for them. Postmates and Instacart made their shoppers commit to being available for multi-hour shifts, rather than having the freedom to accept jobs as and when they want them. This starts to look less and less like “gigs” and more and more like regular old jobs.

So how do we regulate jobs like these? How do you balance the protections and rights we want to guarantee all workers, without killing the opportunities for “gigs” that many find a useful supplement to their income, or a flexible way to work around other commitments?

California is place currently trying to find this balance…. but not very well. Legislation was due to come into force today that would force all Uber and Lyft drivers to be re-classified as employees, but a court has granted an emergency stay this week. In a case of both sides being wrong, the California legislation seems very poorly drafted and overly broad, but equally incredulous is Uber and Lyft’s claims that their drivers are completely independent contractors.

In London, even though taxi-drivers are generally self-employed, the Employment Tribunal found that Uber drivers are not independent contractors. Independent contractors, they reasoned, can negotiate their fee, are free to reject work or set many of their own terms, none of which are options available to Uber drivers. They are therefore more like employees and should get employees rights, like sick pay and minimum wage. Uber are appealing this in the UK Supreme Court at the minute.

Closer to home, our major taxi app Free Now does act more like a simple connector for Taxis, so doesn’t come up against these conflicts, but we do have many “gig economy” workers, especially in food delivery like Deliveroo, Just Eat and Uber Eats. All of these companies engage their riders as self-employed contractors, rather than employees.

Earlier this year the High Court issued a finding that Dominos Pizza delivery drivers are effectively employees, not contractors, even if their contract says otherwise. The High Court’s reasoning was similar to the London tribunal’s. One would imagine they’d apply the same finding to all other gig economy companies, if a case comes before them.

Which would probably mean that services like Deliveroo, Just Eat and Uber Eats won’t be able to operate in Ireland (they already loose a ton of money even without this). And maybe that’s where we’ll end up, that they can’t be viable businesses without exploiting vulnerable workers, in which case we’ll move on without them. But it’s probably worth exploring and writing more explicit regulations for these industries, to try find a middle ground that protects workers, especially those most economically vulnerable, while also giving them and others the benefits offered by more flexible gig work.

The worst case is that we end up where California is this week, with the worst of both worlds.

🔗 Interesting Links

The data team at Yelp have taken all of their usage data and mapped it against Covid-19 outbreaks in the US. Two interesting findings. The first is the specific activities that are linked to outbreaks – “As people began resuming common pre-pandemic activities [….] – specifically, frequenting restaurants, bars and nightlife, and gyms – a clear spike in COVID-19 cases within those locations quickly followed.” Second, even in states without mandated lockdown, people naturally curtailed visits to places where social distancing is harder to enforce. Many temporary closures are becoming permanent too. Link.

You’ve probably seen those fake guardian headlines on Twitter? Some are funny parody, poking fun at the various columnists at the paper, but many are used to stoke outrage at opinions that were never published (see screenshot below). Lots of these are created on a free website, which the Guardian is now suing for copyright infringement. It’s an interesting challenge, balancing parody in freedom of speech, vs. malicious disinformation. I would have thought it would have made more sense as a Trademark / misleading consumers approach, rather than copyright, but I’m very much not a lawyer. Link.

Marietje Schaake, formerly a Dutch MEP, has been joined Stanford University in a role which I saw described as “Europe’s ambassador to Silicon Valley”. Probably a stretch, but the New Yorker have a great profile on her and the lessons America can learn from Europe about regulating tech. “Silicon Valley is not usually seen as a policy hub, but, of course, it is one. I saw there was more Silicon Valley in Europe than Europe in Silicon Valley and I felt maybe this should changeLink.

China’s Baidu (Google equivalent) blanks out parts of its maps, so these intelligent reporters in Buzzfeed used those blanks to find a network of 315 locations that look like active internment camps in XInjiang, presumably for the Uighur muslims. Stellar investigative journalism. Link.

For the stats nerds – How Does Ofqual’s Grading Algorigthm Work? – particularly relevant as our own Department of Education will be doing something similar with Leaving Cert results here. Link

This is a sad read on the real human cost of online conspiracy theories. “A Florida taxi driver, who believed false claims that coronavirus was a hoax, has lost his wife to Covid-19” Link.

The Women Making Conspiracy Theories Beautiful. Link.

It turns out most of Scottish Wikipedia (about 20,000 articles) was written by an American teenager who doesn’t speak Scots, “but simply tries to write in a Scottish accent”. Link.

Peter’s Newsletter 14 – What is QAnon?

August 21, 2020 in Weekly Newsletter

In this week’s issue. A briefing note on QAnon. Plandemic 2 gets cancelled. Fatphobic Instagram? Fact-checking fails again. Earthquake alerts. How the New York Times grew millions of subscribers. Social Signalling in Social Networks and How Tik Tok’s algorithm works.

💡 Ideas

How seriously should we take QAnon? 

Every now and again a new conspiracy theory becomes popular. New world order, flat – earthers, fake moon landing (even fake moon!). Many are interesting and bizarre, but some can also be quite dangerous. The false linking of autism to vaccines, for example, is a very serious national health risk.

So it’s always hard to know how to categorise new conspiracies that emerge. Most are peculiar, but not worth spending much time getting to know, but the dangerous ones are worth studying, especially if you are interested in preventing their spread.

The “QAnon” movement definitely falls into the dangerous category, but it’s very US-centric, so I’m still unsure what hold it has in Ireland. Small but growing, it seems.

What is QAnon? Since 2018 a person or people calling themselves “Q” has been making a series of posts, some containing vaguely worded statements, which conspiracy theorists interpret as predictions. Like a fan looking for meaning in song lyrics, they usually find ways to back-fill “predictions” into real world events, deeming Q to have successfully predicted them.

The rhythm of these predictions (or drops) and the frenzy that follows makes seems to build incredibly strong community between the conspiracy theorists who follow Q. This has seen it grow incredibly fast, with an estimated 1.2m people in one or more QAnon related Facebook groups as of July 2020. It also does what many other conspiracies do – isolates followers from their friends and family in real life. Reddit’s QanonCasualties has some harrowing stories from people who’s loved ones have become obsessed with QAnon.

What do they believe? Well that’s the other interesting thing about QAnon – because it’s directed by a single person/group, it shifts focus every so often and kind of acts as a catch-all of major conspiracies. 5G is mind control. Bill Gates made Covid. You name it, they got it. I’d imagine for followers this keeps the movement compelling and interesting, but also adds up to quite a bleak view of the wider world that you inhabit.

They had two major inflection points recently which significantly drove growth. Most recently Covid and lockdown, but before that, Epstein. To be fair to them, they believed the world is run by a paedofile ring of politicians and billionaires… and they weren’t exactly proven wrong.

So how worried should we be? Globally, quite. The FBI has deemed them a domestic terror threat and they have been behind a huge amount of the anti-mask movement. A Q follower is now a Republican congressional candidate in Georgia and yesterday Trump praised Q followers.

But in Ireland, I think it’s worth just being alert for now. Much of the Q narrative is around Donald Trump and how he is the great saviour, which I just don’t think translates too well here. So we seem relatively safe for now, but it’s worth understanding and keeping an eye on. The very nature of the social internet seems to be leading inevitably towards more dangerous and virulent conspiracy movements emerging.

If you want to read further, The New York Times have a good primer. Aoife Gallagher of the Institute for Strategic Dialogue talked to the Beacon this week about QAnon in Ireland (summary: small but growing). Or if you want to deep-dive, check out Gallagher’s report on “The Genesis of a Conspiracy Theory.”

Or if you’d prefer to listen…. Reply All looked at one QAnon conspiracy in July. The New York Times also did a 9 part series called Rabbithole, which is fantastic.

🔗 Interesting Links

Plandemic was a 26 minute long video posted in May, which pushed several conspiracy theories about Covid and how it was intentionally released by some billionaires. Shared widely by QAnon followers it got several million views across YouTube, Facebook and elsewhere before the platforms stepped in to take it down. The sequel was released this week, but the platforms took more decisive action this time around, blocking it out right from the get go. Read more.

On the flip side of the content filtering debate, Instagram came under criticism from activists this week for removing photos of model Nyome Nicholas – Williams (@curvynyome). The photos seemed to trigger Instagram’s nudity filter, even though they’re very similar to many semi-nude (if that’s the right terminology) photos that haven’t been removed. Nyome and the photographer pointed to the fact that her photos, being of a larger, black woman, being removed while others (of skinny, non-black women) remaining on the site pointed to inherent bias in Instagram – either their AI or policies or both. It seems Instagram listened and have since updated their policies. Read more.

To further exemplify the tight-rope platforms are trying to walk between blocking too much and blocking not enough, Slate recently reported on Google blocking advertising from their articles about racism, incorrectly flagging them as racist. No technology will ever get this balance perfect, so the question we need to ask is which side we should err on? Pulling ads/support from anti-racist content, or allowing ads/support on racist content?

Add this to the growing body of evidence that fact-checking fake news doesn’t work very well. A team in UCD tested some fake news messages on people (e.g. that the COVID tracking app was made by Cambridge Analytica) and it always made people less confident about downloading, even if the story was accompanied by fake news warnings. Link.

Earthquake alerts. Google are planning to enable all android devices to work together to detect synchronous shaking in a specific area, use that to understand that an earthquake is happening, then send early warning alerts to people nearby, hopefully giving them a few seconds to get prepared. Link.

A good exit interview with the CEO of the New York Times. Interesting to see their strong focus on paid subscriber growth and building an agile digital team to deliver that. Link.

I’ll also include two excellent pieces of writing about tech platforms here which I came across this week.

Proof of X” looks at how Social Networks are designed to force people to do things for social signalling, then asks if we can design social networks that encourage people to do things that are good for them.

Eugene Wei’s “Tik Tok and the sorting hat” is a brilliant analysis of the algorithms behind TikTok’s incredible growth in the last 12 months.

Peter’s Newsletter 13 – Do We Like Data Centres?

August 14, 2020 in Weekly Newsletter

In this week’s issue: What should Ireland’s Data Centre strategy be? Is the NDRC value for money? Apple blocks Fortnight. Facebook removes Trump’s video. How we regulate the web. TikTok opens it’s algorithm for regulators. And a podcast recommendation.


How should we feel about Data Centres? This week I wrote a piece about Ireland’s current industrial policy of attracting Data Centres to locate here and ask if they’re really worth the environmental cost? Read more.

Interesting Links

NDRC up for tender. The NDRC was Ireland’s first tech accelerator. It is privately run but Government funded, with a mission to invest in and support new Irish tech companies. Over the last 10 years the company who runs it have received €43m from the state and the results have been fairly middling. They paid their 18 staff €111k on average, and the CEO over €250k, which is wild. The contract is up for tender, which should be announced next month. Link.

Fortnight vs. Apple. Apple’s app store is fairly tightly regulated. They refuse apps for a whole host of reasons, to keep quality high or to remove apps that might be unsafe for users. They also have strict commercial terms, taking a 30% cut on all in-app payments (and companies often have no other options but to pay). This has lead to a few high-profile clashes in recent weeks, with app makers accusing Apple of profiteering. The latest of which is Epic games, makers of Fortnight. Yesterday they gave people the option of paying them directly in-app, using a credit card, bypassing Apple’s in-app payments and their 30% tax, so Apple swiftly removed the wildly popular game from the App Store. If you’re a parent of young kids, you’ve probably already heard about this. Link.

Facebook finally removes a Trump post. The post was removed because it contained false information about COVID (that kids are ‘almost immune’), so it was a pretty clear violation of Facebook policies, but it’s the first time they have removed any of his posts. Link.

Regulate the Web. An excellent overview of the history of how we have built and regulated the internet, and where we might go from here. Link.

Simon McGarr has a piece on Ireland’s Online Safety and Media Regulation Bill and it’s ambitions to create an Online Safety Commissioner tasked with regulating… everything? Link.

TikTok have launched a new Transparency Centre, saying they “believe all companies should disclose their algorithms, moderation policies, and data flows to regulators. We will not wait for regulation to come, but instead TikTok has taken the first step by launching a Transparency and Accountability Center for moderation and data practices. Experts can observe our moderation policies in real-time, as well as examine the actual code that drives our algorithms.” Link.

Podcast recommendation – Wrongfully accused by an algorithm, from The Daily. Link.

Peter’s Newsletter 12 – The Geopolitics of TikTok’s Irish Announcement

August 6, 2020 in Weekly Newsletter

In this week’s newsletter issue – The wider context to TikTok’s big Irish jobs announcement, new stats from Scotland show the bias we’ll see from the Leaving Cert grading algorithm, Australia’s plan to tax big tech to support news media.

TikTok announced that they will be hiring “hundreds” of new jobs in Ireland, investing €420m in the next two years to build their main EU data centre here. Jobs announcements are always welcome (even if they don’t fully deliver on the promise) but it’s worth looking at some of the wider politics at play behind this particular announcement.

TikTok is a social media app owned by the Chinese company Bytedance. From the earliest days, the Bytedance founder has tried his best to position the app as a global tech platform, rather than a Chinese company. Most of their user data is stored on servers are outside China. They hired an American from Disney as their CEO. They’re building up big teams outside the US. They even made TikTok unavailable in China, so videos aren’t subject to the Communist Party’s censorship rules.

All of this to try create the first truly global tech company, based in China. But at the end of the day, geopolitics is catching up. It’s still a Chinese owned company operating under the Chinese legal system, which means it is ultimately answerable to the whims of the CCP. I haven’t seen any nefarious interference in TikTok from China, but it seems like the threat alone is enough to make the US administration nervous about the app’s gaining popularity.

That’s the context within which the Irish data centre was announced this week. “This investment in Ireland” TikTok’s Chief Information Security Officer said, “will create hundreds of new jobs and play a key role in further strengthening the safeguarding and protection of TikTok user data.”

You can see the angle they’re going for here. Even the fact that the “Chief Information Security Officer” is the one they put in the press release is revealing of the strategy.

None of it has helped, it seems, as Trump’s administration is now threatening to ban the app in the US unless it sells to an American company, with Microsoft as the only potential suitor. The Trump administration continues to escalate this even further, announcing a new policy last night that includes a plan to “remove untrusted applications from U.S. mobile app stores.

TikTok’s fate should be interesting to all of us, not just because it’s used by hundreds of thousands of young Irish people, but because it is the first casualty of an escalating digital trade war between the US and China. The position the EU will take in all this, and Ireland as the home of big tech in Europe, is still unclear and certainly one worth watching as it develops.

Leaving Cert Algorithms. Often a good way to think about bias in algorithms is not that they are biased themselves, but that they reveal biases in the underlying system that they’re based on. When you feed in historical data about the way decisions were made in the past, develop an algorithm using that data, but then find the algorithm is making biased predictions, that’s a revelation about bias present in the historical decisions, moreso than the algorithm.

All this to say that we will probably find revealing biases in the calculated grades when the Leaving Cert results get released next month. The grading will done in two phases. First, a student’s grade will be guesstimated by teachers, with oversight by the principal to normalise grades within a School. That seems fine. But then the Department of Education will apply a “National standardisation” of grades, which is basically a predictive algorithm. It looks at the grades the average student in that school got in the past and uses that to predict what an average student would have got in 2020, if the Leaving Cert had gone ahead as normal. If the average grade in a School is higher than predicted, their grades will be adjusted down accordingly.

How might this look in practice? Scotland released their numbers this week. Teachers in the most deprived Schools graded their average student at 85%, but the average mark was brought down to 69%, because kids in deprived schools usually never did that well. Compare that to the least deprived schools, which were graded at 90%, but weighted down to just 84%.

Poor kids can’t get high grades in 2020 because poor kids don’t normally get high grades.

So when the calculated grades get released here and inevitably show massive unfairness similar to Scotland, remember that the unfairness is not in the new grading process, but in the deeply unfair system underneath. The inequality will be no different this year to any other, it will just be quantified and transparent in a way it wasn’t before.

The Business Model of Journalism. In Australia this week, the Australian Competition and Consumer Commission released a draft plan to force payments from digital platforms (Facebook, Google etc.) to pay money to Australian media organisations. It’s an interesting new approach. It claims the problem is an imbalance of power in the bargaining ability of news organisations against tech platforms.

It says that Australian news companies can’t charge Facebook and Google for “deriving a benefit from the ability to make Australian news content available to their users” and blames that on an imbalance in bargaining power.

It’s an odd approach. News articles appearing in Google search results and being shared on Facebook, driving lots of clicks to their websites, seems like something that benefits those media organisations? But the Australian government thinks Google and Facebook should pay for the privilege?

The advertising business model that supported journalism is collapsing and journalism is important to society as a whole, so the case to use a tax to prop up new media, or just directly fund journalism, has merits. Why not just explore that? This feels like a huge number of extra steps because you don’t want to call a tax, a tax.


I send this update in a newsletter every Friday morning. Subscribe for free here.

Peter’s Newsletter 11 – How do we decide which ads to ban?

July 30, 2020 in Essays, Weekly Newsletter

There wasn’t much important news in the world of tech this week, so this week’s newsletter is a deeper piece of analysis on the ASAI’s decision to remove the Tampax commercial, the role of self-regulation and the Facebook Supreme Court which will set up later this year and make lots of decisions like these.

The “Tampons & Tea” Ad

This week the Advertising Standards Authority for Ireland upheld complaints about a tampon commercial which Tampax have been running on Irish TV.

The decision to recommend the ad be removed from air has caused quite a bit of backlash, not least because many saw the ad as genuinely educational. You don’t read this newsletter to hear a tech nerd express opinions on tampon commercials, so I won’t, but I do think it’s interesting to consider how the ASAI made their decision and to ask the more important question… who even are the ASAI?

A total of 84 complaints were made about the advertisement, under 4 broad headings. Under three of the headings, the ASAI didn’t find the complaints adequate. These were “Sexual Innuendo”, “Suitability for Children” and being “Demeaning to Women”.

The only complaints that were upheld were under the heading of “General Offence” and even here the reasoning is peculiar. From my reading of their findings, it seems that, while the ASAU didn’t find it to be offensive, enough people complained that they considered the ad to be breaking the code that “A marketing communication should not bring advertising into disrepute.”

To paraphrase, ‘We didn’t find it offensive, but enough people did, and the rules say not to be offensive’.

In defence of the position, ASAI Chief Executive Orla Twomey said in the last four-and-a-half years there have been only seven adverts that have had 60 or more complaints.

So I guess we know the magic number now. If you want any ad you dislike taken off the air, just get 60 people to fill in an online form.

Who are the ASAI?

Despite what you might first guess, the ASAI is not a government body or publicly funded. It is an industry group which has a self-regulatory code of conduct for all advertisers.

The optimistic view of this setup is that all industry players have a vested interest in keeping the standards in advertising high. If one advertiser tries something crass and eye-grabbing which opportunistically works for them in the short term, but degrades the efficacy of advertising in the long term, this is bad for everybody.

Also, not being defined in legislation gives self-regulatory bodies a bit of nimbleness to adapt as methods and trends change.

The more pessimistic view is that they do “just enough” to keep the worst behaviour at bay, but give government no impetus to setup a public regulator or enforce stricter rules. The main motive of a body like the ASAI is to keep advertising a profitable enterprise over the long term by setting code of conduct for advertisers. Sometimes this profit motive overlaps with the wider goals of society, but sometimes it does not.

This week’s decision is a good example of when those interests can come into conflict. If you’re an industry body charged with keeping a medium profitable, why wouldn’t you ban the one ad a year that generates too much controversy? Assessing it complexly is difficult and erring on the side of conservatism makes sense.

Where as a statutory (non-industry) regulator may have to weigh up a decision more complexly, considering the educational benefit, the wider context of gendered “offence” and the importance of free expression.

This decision is also another great example of how difficult it will be to regulate “political” and “commercial” advertisements as if they are always distinct and separate. Is Always’ “Run Like A Girl” ad commerical or political? Nike’s Colin Kapernick ads? What about Monsanto running ads about the benefits of fertiliser?

This isn’t the ASAI’s first difficulty with this distinction either. In 2018, they refused to hear any complaints about any advertisements in the abortion referendum, which left the digital ads in an un-regulated limbo. In 2016, they found the ads for Eircode to be misleading, it said they were outside its remit because they’re “public broadcasts.” (Which is probably fair – you can’t have an industry body regulating the government?)

Facebook’s Supreme Court

While we’re on the topic of self-regulation, the new Facebook “supreme court” has been established this summer and will soon start hearing cases.

The Facebook Oversight Board, as it’s officially called, will play an interesting role for the company and is an experiment worth watching for anyone interested in the ongoing debate around how we moderate content online. It is made up of some pretty impressive people, mostly former judges and human rights lawyers, and the intention is that it sits separate to Facebook (although funded by it).

Characterizing the FOB is tricky, as Evelyn Douek notes – “It is court-like in that it will hear appeals from and act as a check on Facebook’s policy-formation and enforcement processes and provide public reasons for its decisions. But it will also give policy recommendations, and neither its members nor those who appear before it will be lawyers applying the law. It is a private institution fully of Facebook’s own creation, but it has reasonably robust mechanisms to ensure independence from Facebook, which has put $130 million into a trust intended to fund the FOB for at least two three-year terms. It is a global body, but it would be naïve to think that it will be able to settle global speech norms when different jurisdictions have clashed about these for many decades.”

Facebook currently employees about 35,000 content moderators globally, who make decisions every minute of every day to remove content from the platform based on an ever growing set of company policies. If the removal of an individual piece of content, or a certain type of content, proves very controversial, Facebook can escalate this to the FOB for deliberation.

People who are hoping this might act like the US Supreme Court, handing down binding rulings and setting precedent for Facebook, will be disappointed. This Oversight Board won’t act in that way, and it would probably be naïve to trust that such a system could always work – Facebook can always just choose to ignore a body that it established, if it wishes.

But Evelyn Douek argues, quite compellingly, that this isn’t what optimists should hope for from a body like this. Instead, the benefit should come from the act of public deliberation itself, rather than just the final ruling. Forcing Facebook to explain why they removed content, and defend the logic publicly, should greatly enhance the public debate around these issues and also cause Facebook to more carefully consider each internal policy they implement, knowing that they may one day have to defend it publicly.

The hope is that the dialogue between the FOB and Facebook, through being forced to make arguments in cases and publicly respond to FOB recommendations, will finally ventilate the reasons behind why Facebook makes the decisions that it does by forcing Facebook to justify them. This process itself will hopefully improve decision-making, but at the very least it will provide a level of transparency and accountability that is currently sorely lacking. To those from the United States, the paradigm “strong-form” judicial review jurisdiction, this might seem feeble. But many other jurisdictions have a version of this dialogic “weak-form” review, and it often turns out to be much stronger in practice than it appears in theory.

As we’ve seen this week here in Ireland, self-regulation has its own problems. As Douek notes, “it is unsatisfactory for private, profit-driven platforms to be making these decisions unilaterally and without any accountability.

“On the other hand, heavy-handed government involvement in speech regulation is always suspect, and the cure to our current woes should not be worse than the disease. The FOB is therefore an effort to find a third, least-worst option.”

Similar to the industry that the ASAI represents, much of Facebook’s business model would just be simpler if it was an uncontroversial space to sell people’s attention to advertisers, so why not delegate some of these decisions away?

These are issues that will only become more and more prevalent in the coming decade, as the public sphere shifts from broadcast and print to digital, so every new experiment is worth watching closely. I’ll be watching with skepticism and a dash of hope.

Peter’s Newsletter 10 – The TikTok Trade War

July 23, 2020 in Weekly Newsletter

I send this update in a newsletter every Friday morning. Subscribe for free here.

In this week’s issue: 1) The HSE Covid Tracker accused of tracking private info, 2) Why countries are banning Chinese tech, like Huawei and TikTok, 3) A YouTube algorithm art project, 4) Crowdsourcing maps of trees in Dublin, 5) Which counties pay the most income tax? 6) A new Irish game about growing up and coming out in Ireland, 7) Ireland shares it’s Covid App with the world and 8) Twitter mutes Q-Anon.

1) Is The Covid App Violating Privacy? (No, but several news stories reported that it does)

If you have the Google version of Android on your phone (as 99% of Android users in Ireland do) and want to install apps on the phone, you’ve probably created a Google Play account. If you did (and everyone does), then your phone pings Google HQ many times throughout the day with info about where your phone is and what its doing.

Google also developed the Exposure Notification Framework, which Covid Tracking apps in many countries, including Ireland, are built upon. If you’re using the Covid Tracking app on Android, you’re also a Google Play user and Google has your data.

A research paper published this week from Trinity points this out – that the Covid Tracker App works like all other apps. The paper asks, given that the aim is universal adoption, if Google should also provide a non-Google solution. They say this one type of app should have the option of working unlike all other apps, and be usable while avoiding Google’s data mining infrastructure.

That’s fair question to ask and an interesting angle for privacy campaigners to use to pushback on Goolge’s data hungry policies. But unfortunately it’s being reported on in some cases as “The Covid App shares all your data with Google”, which is a pretty unfair characterisation and risks undermining public confidence in the app.

2) TikTok and the Chinese Trade War. Tech is the arena where the growing tensions between China and the rest of the world are playing out. The US first, and now the UK, are starting to mandate the removal of Huawei technology from their mobile networks. India banned TikTok a few weeks ago, and there’s mumblings from the US about doing the same.

The core issue is the question most democracies will have to grapple with when thinking about how they engage with these new, large Chinese tech firms. How dependent do you want to become on any company that, at any time, could be forced to carry out the foreign policy aims of the Chinese Communist Party? I certainly don’t know the answers, but here’s a few thoughts:

  • For the last 20 years, the internet and global tech landscape has been shaped by the cultural norms and legal system of one country – the US (and in particular, Silicon Valley). This won’t be the case for the next 20 years, the US will be joined by many more, in particular the EU and China.
  • A common concern is phrased as – “What if” the CCP demand that TikTok’s algorithm promote videos in favour of the CCP foreign policies (burying videos about Uighur genocide, or Hong-Kong democracy, or disrupting an EU election) on a platform that the majority of young Irish people watch every day?
  • This feels new to Americans, but the rest of us are used to using tech built in foreign countries. The rest of us have been very worried about the impact of selling political reach to the highest bidder on US tech platforms, for example.
  • If the EU is comfortable saying our data shouldn’t be transferred to the US because of their surveillance laws, is blocking Huawei or TikTok because of CCP policy much different?

I think The Economist had the best take on this with their recent cover “Trade without trust”. An escalating tit-for-tat hurts everyone. We need to find a way for the EU and US to do trade with an authoritarian country that they can’t trust but can’t ignore. Link

3) Show me your YouTube. This cool art project lets you walk a mile in someone else’s digital shoes, by experiencing their YouTube recommendations. See what the YouTube homepage looks like for a “liberal, conservative, climate denier” and others. Link.

4) Mapping Trees. Dublin City Council and the National Tree Council have launched a project to crowdsource the mapping of every tree in the city. Read more http://www.dublincity.ie/dublin-tree-map

5) If Found…. Is a new, award winning game (on iOS and PC) about coming of age in Ireland in the 90s. As game dev becomes more accessible, it’s great to see more small teams in Ireland experiment with it as a story telling and art form. Link to the game. Dublin Inquirer writeup.

6) Income Tax & Jobs Dashboard. The PBO have produced another cool dashboard, this time an interactive map of disposable income, taxes and employment levels per county. Link.

7) Covid Green. Ireland have donated the source code of our new Covid Tracker app to an open source foundation, so that it can be used by other countries who need it. Great to see. Read more.

8) Q-Anon No More. Twitter has announced they are suppressing Q-Anon related content. This is a big move and really good to see. If you’re not aware of Q-Anon, it’s a new but fairly destructive conspiracy theory that’s doing the rounds. In the same space as 9/11 denial and New World Order. These conspiracy theories have incredibly destructive mental wellbeing and social isolation impacts on those who adopt them, so moves like this are great to see Link.

If you’d like to receive an email like this every week, you can subscribe for free here.