A 5-post collection

A world without teachers: YouTube, incentives and the problem with self-regulating platforms

Ally Law, a 20 year old YouTuber, breaks into the National Theatre (Link)

YouTube has been in the news a lot recently. Whether it’s Logan Paul filming a suicide victim in a Japanese forest or Ally Law breaking into the Celebrity Big Brother house, it’s undeniable that YouTube - along with many self-regulating platforms - has a significant problem. YouTube, and platforms like it, make more money if viewers are engaged for longer. The more eyeballs they control, the more ads they can sell. For top creators - typically young, creative and hungry for fame - this is important. View count and engagement are the critical metrics that can make the difference between a video with revenues of a few hundred dollars or many thousands: universal indifference or global fame.

Incentives & The rule of the playground

There is one deep fear among these young YouTubers: having a video demonetized. If YouTube believes, for any reason, that the content you’re producing is not advertiser friendly then it will demonetized your video and prevent you earning money from it. Reasons why a video will be demonetized include:

Discussion of controversial issues and sensitive events, harmful or dangerous acts .. [or] pranks involving sexual harassment or humiliation
Read the full list here

This content is allowed on YouTube, it just can’t be monetized by creators. The decisions about what is appropriate and what is not are taken initially by pieces of software and, if escalated, by YouTube’s (small) team of moderators. These moderators are judge and jury; though unlike most judges the moderators are incentivised to keep content monetized that’s performing well - it’s YouTube’s revenue after all (YouTube splits the ad revenue 40:60 with creators).

From a regulation standpoint the incentives are clear: create content that garners the most number of views, for the longest possible time, whilst staying within loose regulations that primarily include a prohibition on sexual content and swearing. Beyond that, anything goes and anything is permissible.

Think about the sort of events in your life that encourage people to linger and observe: traffic slowing to observe a traffic accident on the opposite side of the road, a fight breaking out on the adjacent platform, a child at school jumping from a roof. These forms of entertainment are some of the most unsophisticated but effective. In YouTube’s world this is even better: content like this crosses cultural boundaries, language and more.

The YouTuber Jake Paul's most popular videos

If you needed any indication as to what content performs the best, consider Jake Paul, Logan’s 21 year old brother and amasser of 13 million subscribers. Some of the most popular videos include: “I CHEATED ON MY WIFE PRANK (she freaked out)” with 21 million views, and “RANDOM TATTOO SPIN WHEEL GAME (You Spin It, You Get It…)” with 18 million views - described by Jake Paul as “ONE OF THE MOST SAVAGE THINGS WE'VE EVER DONE”. The content that thrives on YouTube is the same content that thrives in the playground: bullying and sensational. Unlike the playground though, in YouTube’s world there aren’t any teachers to regulate what goes on - Jake Paul’s bulling behaviour can continue long after school (and make him a millionaire at the same time).

And Jake Paul really is a bully. The Martinez Twins who used to live in the "Team 10" house with Jake Paul managed to leave and posted a video about the abuse they suffered. Paul left them afraid to sleep by insisting that they left their door open at night and "pranking" them awake with "funny things" like a tazer. This wasn't new to Jake Paul:

A world without teachers

In a sense this problem has existed in the media since its inception. The Daily Mail’s “side bar of shame” is but one example of a publication driven by sensation and views. The one difference is that mainstream media does have at least some regulation. YouTube’s chief business officer said today that YouTube was "different" from traditional media outlets as it doesn’t have the same "editorial hand”. This is absolutely not the case - a moderator making a decision to allow a piece of content based on guidelines is exactly the same decision making process as an editor in a traditional newspaper; the only difference is that YouTube refuses to take this heavy and important responsibility. Why? Because admitting that they're a publisher carries with it a huge cost. 300 hours of content are uploaded to YouTube every minute - 432,000 hours every day. If you assume moderators work an 8 hour day YouTube would need 54,000 moderators just to watch the content, let alone make any decisions. To put that in perspective, Alphabet - YouTube’s parent company - currently employs around 72,000 people globally.

This raises two fundamental questions:

1. Can AI seriously replace the moderation of human editors?

Platforms like YouTube argue that they don’t need a large army of moderators because AI can do the job for them. Any mistakes made so far are because the AI hasn’t been trained properly. In my view this fails to take into account the subjective and nuanced view of an editor - the decisions made often aren’t clear cut and can’t be based on previous data. Context is critical for decision making - racism or abuse might be subliminal - something a human can pick up on but a machine never could. If this is the case, is it acceptable for YouTube to leave 99% of content unwatched? Publishing what could be incendiary content in the dark and waiting for enough viewers to flag it.

2. Should YouTube be the arbiter of what is acceptable content?

YouTube has a clear motive: increase viewership whilst not going too far so that advertisers will stop advertising. However, publishers have responsibilities as members of society. The ideas, thoughts and content created by YouTube permeate the world. Free speech is guaranteed, of course, but if freedom of speech is YouTube’s defence then they must also bear the true responsibility of free speech. As Areeq Chowdhury the Chief Executive of WebRootsUK said recently:

"Freedom of speech doesn't mean you can simply say whatever the f*ck you want about something, without there being consequences."

Promoting content like Ally Law’s which involves breaking the law again and again is YouTube's responsibility too. It shouldn’t be possible for a platform that publishes videos that involve breaking the law to take absolutely no responsibility whatsoever. Tacit support is still support, and simply closing your eyes doesn’t remove your responsibility.


Tech companies are ripping society apart (and this is only the beginning)


We are in the grip of a technology revolution and, as many are quick to point out, we’ve been here before. The industrial revolution in the 1800s saw a fundamental shift in the way that society operated, but despite the disruption to traditional livelihoods broad ideas about capitalism and the economy remained true. This analogy is used to reassure anyone worried about robots taking jobs & technology companies running rampant. Society will adapt and things will return to some sort of equilibrium.

I believe that this view is misguided. There are elements of the current technology revolution that are leading us towards the most unequal dystopia imaginable. These aren’t unintended consequences that can be cleaned up later like the smog rising from industrial factories: these are paradigms embedded directly in the current technology revolution that, left unrestrained, will corrode the very fabric of society.

1. Today’s technology creates vast wealth for a tiny few

Technology’s aim is to reduce cost. A store front used to cost money; with the web it’s now free. A map used to cost money to print; now maps are available on your smartphone for free. This has two effects: creating more abundance for consumers, and more wealth for technology companies. An online shop is cheaper than a physical shop because an online shop employs no one and costs no rent. In addition to this, a physical shop can only service a few dozen customers whilst a digital shop can service millions. Essentially this means that for a lower cost a shop owner can generate many millions of times more revenue and profit - with very little associated costs.


As more and more jobs are replaced this shop analogy will hold true for every industry we interact with: abundance for many, profit for few. This has started happening already: since 2003 the top 1% of wage earners have seen their annual household income rocket in comparison with other groups. If this wasn’t bad enough each year we will see greater increases for a smaller percentage. 1% will become 0.9, 0.8 and so on until the cost base of operating most industries is so low that just a handful of people will be profiting.

2. Today’s technology replaces good jobs and replaces them with terrible & insecure jobs

As technology companies there is job growth in some areas. Amazon is now one of the largest employers in the US. This is a problem, though, as the vast majority of Amazon’s (and most technology company’s) jobs involve low skilled, low paid work, in Amazon’s case inside fulfilment centres. In early 2017 their CFO said that “The headcount [we’re adding] is predominant[ly] still the headcount in our fulfillment area.” As artificial intelligence capabilities increase these jobs will be replaced (one of the key problems now is that robots can’t grip very well) with equally low skilled jobs like cleaning and delivering, and in the meantime there is no opportunity for employees to move up within Amazon. There is a clear divide between the engineering class and the lower skilled work done by the vast majority of employees.

The fulfillment centre of the future

This is compounded by the fact that there is a clear divide among demographics: almost all of the high paid engineering work is done by white or asian middle and upper class men. Technology companies typically generate a tiny number of high skilled jobs that often go to well educated people from a privileged background. These high skilled jobs will stay but the low paid jobs will go. As low skilled worker’s wages stagnate because of an increase in supply due to automation more money will be freed up to pay high skilled engineers. Again, this is already happening: wage growth amongst the top 10% has increased substantially in recent years yet in the bottom 10% it has actually decreased.


3. Today’s Technology is almost impossible to tax

One would hope that the societal damage that is being caused by technology companies would be offset by an increase in tax payments. Ironically though technology companies pay even less then their historical counterparts.


Through a combination of tax loopholes (operating in the UK, based in Dublin, Luxembourg or similar) and writing off any profit as expansion cost whilst delivering the actual value through an increase in stock price, technology companies are simply not paying enough tax.

The disruption to society can be solved through retraining and reskilling but these initiatives cost vast amounts of money.

4. Workers can’t be protected from technology by traditional institutions


It used to be trade unions that lobbied for workers and protected them, to some extent, against the excesses of capitalism. Unions today are becoming more and more irrelevant. Many jobs today are insecure and irregular: people have portfolio careers and so don’t neatly fit into the union mould. Collective bargaining works less well when there’s an oversupply of people ready to pick up where you left off, and the lobbying power of technology companies in government prevents genuinely progressive legislation from being written.

There are signs that the traditional union model is being replaced by something more flexible but we’re still a long way off. While unions organise, technology companies have taken the initiative and offered readily available work - as long as workers accept the insecurity (often called “freedom” by tech companies).

What next?

Society gives business a licence to operate. A genuine conversation needs to be had about the impact of today’s technology on the society of the future. There are real dangers that left unregulated technology will push society to a point of no return, increasing radical and potentially unsavoury political actors to take advantage of this dissatisfaction. There is a chance to avert this nightmarish future, but we need to act now.


Taxing data

Recently Google has made a large push towards their photo sharing app, Google Photos. Offering unlimited high-resolution storage for anyone who buys a Pixel phone, and standard resolution for anyone who signs up, it seems like a deal too good to be true. Whilst signing up is easy, removing yourself from the service (as I tried to do a few months ago) is much harder. Photos can only be deleted in batches of 500, and the process will fail four out of five times. It’s the sort of dark user experience that you can see in other products, but it is a reminder of an important point: Google is a data collection company, and its value is dependent upon the quantity and quality of data that it can collect.

The problem with taxing technology

Historically tax has been used to collect revenue for the state by extracting a proportion of the value that businesses or individuals create. Taxation from Biblical times is described as

"When the crop comes in, give a fifth of it to Pharaoh.

Genesis, Chapter 47, verse 24

In this case, the taxation system is quite simple: whatever value you create (i.e crops), give 20% of it to the Pharaoh. Since biblical times this principle of extracting value from producers has been extended to value added tax (VAT), where every transaction that creates some element of value is taxed at a percentage. The bottom line of a capitalist taxation system is that where value is created it is only fair to take a proportion of that transaction as tax.

When it comes to technology companies, the value almost always sits in the data that they own. Citymapper, a UK based transport company is valued at over £250 million, despite the fact that it doesn’t make a profit. What it does have, though, is huge amounts of data about transport patterns in cities all over the world - often much more accurate data that municipal transport companies themselves have access too. It’s a similar story with Uber, which recently launched Uber Movement, a service demonstrating just how valuable their data collection of journey times has been.

Citymapper recently launched a pop-up bus service to demonstrate the power of the data they have access to

The issue with taxing these companies is that though an organisation like Citymapper is generating substantial value with the data that it’s collecting, there is no concrete monetary figure associated with it. Therefore, for most fast-growing startups the core of their value goes completely untaxed.

Where value goes, tax should follow

Over the past few years, the regulation on the collection and processing of data has become stricter. In order to collect data of any kind in the UK, you need to register with the ICO, and soon companies will have to comply with strict security legislation brought about by GDPR. Despite this regulation of data, though, there has been absolutely no attempt to tax the collection or processing of data.

Google’s search algorithm is improved with every search that is made. Though there is a monetary cost of providing the service, the benefit of having the most predictive algorithm is clear to be seen. Terms of service across any tech company that you care to mention includes clauses about using consumer data to improve the service for others. This may seem like an altruistic motivation, but in reality, it is anything but. Taking the insights from Google search and integrating it into Google Home, for example, allows Google to create a much more valuable product. Again, this value is completely untaxed.

Government & regulators are unprepared

The shape of the global economy has been dramatically transformed by technology, and yet the regulatory and taxation system has remained relatively static. As technology companies experiment with new ways to create value it’s the responsibility of government to experiment with new ways of taxing them. Sadly the exact opposite has happened. HMRC recently removed the flat rate VAT scheme benefit for what they call “limited cost business.” Typically these businesses are tech companies whose only capital expenditure are laptops at the beginning of the year. From then on they operate with limited costs - but this doesn’t mean that they’re not creating value.

Simply closing the option for companies to claim back flat rate VAT doesn’t solve the problem, it only seeks to demonstrate the difficulties of creating a fair taxation system where most of the value is invisible when measured with traditional metrics. We need a radical rethink of the way that value is measured and tax is administered, and this has to start with a technology first approach.


Too much content, not enough curation

Why online education is failing

Every school or university has a library. Some universities have dozens of libraries. Everyone recognises, though, that simply having access to information stored in a library is not enough to give someone a good education. Teachers in schools and universities guide, curate and inspire students to find the right information and to present it back in a coherent fashion. Teaching a way of thinking, rather than just studying facts, gives students the ability to learn for themselves and become adaptable and critical of new knowledge and ideas. With no guidance, the only way of studying books becomes to start at “A” and work your way through to “Z”.

A university with no teachers

Technology gives anyone with an internet connection free access to information. The Web contains more content than any library, and more information than any single human being could ever know. Yet like a university with no teachers, there is no one to guide or curate this content. For people who want an education, there is a core component of knowledge missing from the web: how to learn. For those who know how to find relevant information online and utilise it, the web can play an important role in educating and improving their intellectual capital. But there is a key precondition: knowing where to look.

Open source teaching

To counteract this problem many universities are now opening up their courses to outside observers. What used to be restricted to a privileged few is now available for the masses. Want to learn about photography? Take a free course from Harvard. Need to brush up on Philosophy? Go to Edinburgh University’s website. Although these courses are free, they ironically ignore what is best about the web: that a lot of the content and information is out there already. Rather than recreating much of the content that has been developed already, a teacher might just select the best resources relevant for the student. Teaching a way of thinking is not teaching the facts.

A new model

What does a teacher fundamentally do? They help you achieve goals; developing your way of thinking and challenging you. Content alone can’t do this, yet technology might be able to. Developing responsive and personalised challenges and tasks could help a student to reach their goals in a scaled and open way. Importantly, the learning is goal focussed, allowing someone to apply their knowledge to a new job or opportunity. Though someone may not know what they need to learn, they will certainly know where they want to get to in life. In this way curation on the web could help them get there; turning the content on the web to good use and democratising the knowledge for people who don't know where to start.