Tuesday, 29 May 2018

The Cambridge Analytica scandal isn’t a scandal, it’s a business model

Cambridge Analytica recently closed down following a joint investigation from Channel 4 and The Observer which exposed that the company had been collecting Facebook user data without consent and targeting messages during the 2016 European Union referendum.

Since then, founder and CEO of Facebook, Mark Zuckerberg has faced questions from the US Senate; which mostly consisted of people who barely knew what they were talking about, piecing together semi-literate, broad questions which were mostly irrelevant. And, the EU Parliament; which entailed 75 minutes’ worth of very specific and nuanced questions followed by Zuckerberg speaking vaguely about general themes that came up, rendering both of these sessions rather pointless. Facebook needs proper scrutiny and despite what Zuckerberg thinks, the technology industry cannot regulate itself. This is evident by the fact that the recent Cambridge Analytica scandal was not caused because Facebook enabled it - rather, it was caused because they encouraged it.

Facebook’s business model is rooted in the notion that companies pay to target adverts based on user data that Facebook has collected. Despite what many might think, the customers aren’t the users, the customers are the organisations and businesses willing to pay for user data. What Cambridge Analytica did was simply an extension of this; they found a loophole.

Let's be clear, targeting political adverts based on user behaviour and interests is not illegal. If it was, Facebook would not be the tech behemoth it is today. They’ve been reasonably transparent about this too as they state exactly what data they collect and how they use it in their Privacy Policy.

Critics enjoy pointing that this is unethical as nobody reads privacy policies because they are too long - as The Wall Street journal pointed out, revised and condensed policies from 30 of the top apps still span the length of an American football field. They could shorten these policies however realistically, unless the whole policy can be summarised in less than 100 words, it is very unlikely that the majority of users are going to read it; a task that would be virtually impossible.

Facebook collecting data and aggregating it for advertising use is how they make their money. Yes, Cambridge Analytica went above and beyond to exploit this but ultimately, it’s not too different from any Facebook targeted advertising.

Saturday, 19 May 2018

When news isn't news

Modern news apps aren’t as ground-breaking as Google and Apple might think

At I/O 2018, Google announced they are relaunching their news app so it brings users a customised feed of stories based on their interests and what they've previously read. In short, this is Google's take on Apple News. These news apps, which are basically glorified RSS feeds, are well-intentioned however ultimately problematic. Technology promised to eliminate some of the issues associated with old media, however I fear that they've only made them worse.

The main issue is that letting an algorithm pick news for you based on interests you've told it and your reading history is not a good way to digest news. Google has proved their algorithms for generating content don't really work – as demonstrated by the YouTube Trending page where fake news, violent videos, and copyrighted material often surfaces to the top. To be fair, this isn't a problem limited to Google. Facebook makes a hash of providing news too. They spent the whole of 2016 doing nothing about the unchecked saturation of fake news circulating around their platform.

Secondly, displaying stories based on previous reads does not mean they are the pieces you should be digesting. Google says, through “reinforced learning,” the app will learn what topics and publishers you like. Apparently "the more you use it, the better it gets." However most people are drawn in by clickbait and non-stories with glorified thumbnails. Fall for them three or four times and presumably this means Google will start thinking that you have an appetite for pointless listicles and trivial news, so you will be shown more.

Thankfully, there is a simple fix to these problems: editors.

Call me old fashioned but what is wrong with allowing actual human beings to decide what are the most important issues of the day and giving stories some gravitas by putting them in the public eye? This would serve readers better too as it would diversify the topics that appear in the feed.

The obvious opposition to this idea is the fact that editors have biases. Of course they do. But, as do algorithms. It has been proved numerous times over the past decade that algorithms reflect the people who write them so they're not truly impartial either. At least with a group of editors, there is someone to hold accountable if you don't think the news is neutral.

In a society where there is now a large need for views to be challenged and news not to simply reinforce a person’s 'confirmation bias', Google News is doing little to address the challenges. In an ideal world, Coffee House blogs from The Spectator should be put in front of liberals and ‘Comment is Free’ think-pieces from The Guardian should be exposed to conservatives. Instead, it would appear that Google is keener to associate their news efforts with buzzwords like 'machine learning' and 'AI' in order to satisfy investors - rather than creating an information tool that would benefit society.

This is nothing new though. It's unlikely that someone on the political left would go out and buy The Daily Mail or someone on the right The Observer. However, technology has the ability to bring people more balanced news and to expose people to analysis that they might not agree with. Moreover, the companies which are providing platforms for news (Google, Apple and Facebook) have a duty not just to host the content but to prioritise the content which is actually factually correct. Simple as it may sound, this is something an algorithm currently struggles to do.

In Google's defence, they do try and partly address this problem. The news app has a tab called 'Headlines' which Google claims will help readers "understand the full story." On Stage at I/O, Trystan Upstill said, "Today, it takes a lot of work a lot of work to broaden your point of view and understand a story in depth. With Google News we set out to make that effortless." Cut through the waffle and fundamentally this is just a feed of top stories from a wide variety of trusted sources.

Inside of this, they’ve created a 'Full Coverage' feature which will allow users to see how a story is being covered from a variety of sources - including fact checking sites. Google says that everyone will see the same content in Full Coverage. "It’s an unfiltered view of events from a range of trusted news sources."

So why not build the whole app with this foundation, rather than hiding it in a niche corner of the app? When you open the app, you still land on the 'For You' tab which serves you a plate of sources you like the taste of.

Although Google – and others – have the capacity to incite positive change to the media landscape and democratise information to provide the world with more balanced news, Google News has the same problems as old news media, but in a modern format and with “light, easy, fast, and fun design choices.” Thanks for nothing, Google.

Thursday, 17 May 2018

Google's quest to end real life human interaction

Why talk to another person when you can talk to Artificial Intelligence?

Google I/O 2018 was where Duplex was announced
At Google’s developer conference earlier this month, the tech giant announced a new feature in Google Assistant called ‘Duplex’ which will make calls to organisations and businesses on your behalf. “Hey Google, book me an appointment at the barbers” - and voilĂ ! Google will make that very simple call that you are apparently too busy to make.

I am sceptical of this new gimmick. Ethically, Google Duplex will need to start every call by declaring that it is a bot. Legally, I don't see how this gets around the two-party consent needed in some US states for recordings to be made (as all the processing will be done on Google’s servers). And seriously, why does the voice speak with fillers like “urm” and “hmm”? Google has a lot of questions to answer.

Say what?
Even though CEO of Google Sundar Pichai says, “Duplex can understand the nuances of conversation,” I suspect the prospect of tech like this working seamlessly is still a long way off. When I was watching the keynote on YouTube (a Google company), I couldn’t help noticing that the automatically generated captions were still nonsensical drivel – proving that text-to-speech and vice-versa technology is still more of a novelty than a reliable tool.

In the demo, they got Google Assistant to call a 'real' hair salon and book an appointment. However the conversation still seemed rather simulated. I wonder how Duplex will respond to disgruntled, underpaid restaurant workers with a strong accents. Will Duplex give as good as it gets when called a “wee bint”?

Thinking about the future of this product, offloading scheduling calls that you don’t want to make to restaurants, masseuses, and alike are just the beginning. How long after the launch of Duplex will I be able to say, “Hey Google, catch-up with my best friend”? Then for it to call them; outputting extracts of my voice so it can murmur along and ask broad follow-up questions as they drone on and on about their banal endeavours in the club last night.

Tech giants like Google always appear to have altruistic intentions on the surface, however I can’t see many realistic long-term advantages of Duplex. It will only decrease actual human interaction and make the few interactions we do have with strangers more negative. What kind of message are you sending to the kabab shop or dry-cleaners if you get Duplex to call them rather than making the simple call yourself? You’re basically asking for your someone to gob in your doner meat or lose your knickers, Mrs I’m-too-important-to-deal-with-you-little-people.

I’m labouring the point here, but Google is making the large assumption that people will want to interact with bots. I know I certainly don’t. So, even though calls to businesses are often mundane, cherish these little pieces of actual human interaction whilst you can.

Sunday, 13 May 2018

Interview with Queen Mary's Principal for The Print

Last month, I interviewed President and Principal of Queen Mary Colin Bailey for The Print. We discussed the higher education sector, student safety in London, his vision for Queen Mary and much more. Listen below or on iTunes. Alternative, search 'The Print Podcast' wherever you get your podcasts.

Questions were by myself and Sophie Mitchell, with additional help from Queen Mary students who submitted questions via our web form.

Apologies for the poor sound quality at times; the interview its self was recorded using my phone's microphone.

I'd love to hear your feedback and suggestions for future guests who would make good interviews (preferably people who do something that is of interest to students). So, feel free to Tweet at me. In the meantime, happy listening.