Please review and provide feedback on my first blog post

Hey guys,

It's a bit lengthy, but hopefully interesting.


Weaponizing Artificial Intelligence – How impressions win elections.

Facebook has been in the news quite a lot lately, and not for good publicity. Its consequently made some drastic interface changes, affecting the entire digital marketing industry. I’ve been following this for a while now, hence why I wrote this post explaining what’s been happening, what actions Facebook has taken, and what the potential impact could be.

Facebook recently announced that to increase transparency, anyone will be able to view all the paid ads a page is running, as well as require political advertisers to verify their identity before running ads. This will have a huge effect on advertisers, as it means that competitors can easily see: all the ads a page has ran for the last 4 years, total amount spent, number of impressions gained, as well as demographic targeting. Before I discuss the potential implications of this huge game changer, I’ll explain why this happened, what else has been going on, and how, ultimately, it all comes down to one thing.

Facebook decided to take this step following the Russian Ads scandal, which involved Fake Facebook accounts, apparently from Russia, spending over $100,000 on ads aiming to deceive and influence voters in the Trump Vs Hillary Election. The objective was to promote Donald Trump and divide Americans over some of the nation’s most sensitive issues.

Below, is one of the many ads made by the Russian political-militants, which blatantly calls for the removal of Hillary Clinton from the presidential ballot. As you can see, it has quite a lot of engagement, and overall, it’s reported that the ads reached over 125 million Americans.




The issue first came to light in July, but Facebook declared no foul play. However, after mounting concern, it reinvestigated the matter and confirmed that Russian groups have indeed used the social media platform to spread misinformation leading up to the election.

The reports were slightly varied depending on the source, but one thing was common; The creation, delivery, and overall management of the campaigns was sophisticated, professional, and well thought out. From picking issues that were closely linked to Americans’ hearts, such as the Second Amendment, and immigration and gun laws, to effectively using targeting to identify relevant people who could be swayed and influenced. This makes one thing clear, the strategy was well planned and executed - so who ever created the ads, knew exactly what they were doing.

If that wasn’t enough, in September, Facebook disabled the detailed targeting feature after reports surfaced about hate groups using it maliciously to find others with extreme views. This had a huge impact on nearly all detailed targeting ads, which saw a sudden performance dip lasting over 2 weeks. Since then, Facebook has been a lot firmer with who can be targeted in terms of job titles, employers, interests and so on. And in what goes against any self-respecting tech firm’s ethos, Facebook decided that to negate this from happening again, it will give its algorithm less control, and use actual humans to manage this feature – which speaks volumes. Now, although this isn’t directly related to the overall consensus of this post, it’s still relevant and important to include as a recent event that strengthens the emerging theme.

This leads me to the juiciest part. As I mentioned, I’ve been following this for a while, and after reading the various stories, I became intrigued. I started searching around for more info, and it wasn’t long before I came across what makes both the above events seem minute in comparisons.

The original article I read was by Carole Cadwalladr, a seasoned journalist for the Guardian, who’s been shortlisted for her journalism and book. She referred to it as the weirdest story is has ever researched, also adding, one of the most profoundly unsettling scandal of our time.

It all starts with a AggregateIQ, an obscure web analytics company in Canada, which was approached by the official Brexit leave campaign, Vote Leave, amongst four others, who spend over £4 million with the firm in aiding its campaign promotion activities during the referendum. After being baffled by how this unknown data mining and analysis agency played such a pivotal role in Brexit, the journalist did some digging around. She found that it’s connected to Cambridge Analytica, another data analysis agency, owned by Robert Mercer. Robert Mercer is a hedge fund billionaire, and also happens to be one of the most talented data scientist in the world, mostly for his pioneering work with AI.

Cambridge Analytica essentially specialises in furthering political campaigns, although they’ll take on any client, be it political or commercial, if you pay them enough. But this isn’t about money, it’s about power. The power of knowledge from data. And the more data you have the more power you hold.

One of the current hot topics right now is the power Facebook’s algorithm has from all the data it gathers on its users. It is said that in most cases, it knows you better than your closest family and friends. It does this through analysing your activities on the platform, then building a profile based of your 5 major personality traits. It then decides what to show you according to what you like and therefore likely to interact with.

Using his fortune and advanced underspending of data manipulation, Mercer, or Cambridge Analytica, was able to take this a step further. The strategy was simple. Mine as much data as possible, from any source available. The agency did this by buying all the third-party data they could get their hands on. The company bought consumer datasets on everything from magazine subscriptions to airline travel, then anchored the data. It matched all the information back to Facebook profiles through phone numbers and email addresses, with the goal of psyching voters.

In doing so, Cambridge Analytica could not only gauge who was most likely to be influenceable according to their personality traits, but which issues were most likely to capture their attention and guarantee engagement. This created the latest buzzword, from big daddy data and all-seeing mother Facebook having a baby – Welcome: Micro Targeting.

To put this into perspective, the agency married big data and Facebook, then served users that were rated neurotic with images of immigrants flooding the country, to play on their anxiety. Overall, the key was to find emotional triggers for each user, and use highly tailored ads that would get them to act out on that fear.

Through this, the objection of scaring the public into voting leave would be achieved. Again, just to reiterate the brilliant simplicity behind this strategy; statically, the EU Referendum came down to about 600,000 people. That’s less than 2% of registered voters, who would decide whether Britain leaves or stays in the EU. So Cambridge Analytica, or more accurately, Robert Mercer, only had to influence a tiny fraction of voters, by identifying and targeting micro segments who were more likely to be easily influenced.

Conclusions.

It’s no surprise that Facebook has taken this action to improve transparency following the Russian Ads scandal, especially for political advertisers. But what does this mean for us marketers? Personally, I think that knowing spend and impressions levels won’t make a huge difference. After all, it’s expected that the bigger rivals will have more reach and impressions from bigger budgets. I guess now we will be able to see exactly who spends the most.

Being able to see what kind of creatives you’re up against, but more importantly, the targeting strategy of your competitors, is where the true insights are. This will enable marketers to get a good idea of how competitive a market is. This way we can be smarter by deciding whether the market share is viable to chase or not, rather than look at CPMs increase over time and put it down to a mixture of competitors and FB rising prices for ad space to increase its quarterly profits (I’m sure that will still happen).

I think it’s a good change, simply for the fact that it provides us with more insights to factor into decision making and resource allocation.

Looking at the bigger picture in terms of the emerging theme, it’s clear that people, parties or groups are using Facebook for their malevolent agenda. Whether it’s influencing an election or referendum, to what sounds like building an army of extremists. Facebook’s response has been just as extreme. The platform isn’t just adding extra security measures, but also implementing tighter controls, and basically saying that it doesn’t trust its own algorithm, or how people use it.

Overall, people tend to be on one side of the fence or the other. Some people are praising this new technology Cambridge Analytica has created, saying that we need to learn from their micro targeting strategy and use it commercially. After all, you could say that it’s the perfect case of serving the right ad to the right user at the right time. However, some of the tabloids refer to it as brainwashing and distortion on a national scare. Incidentally, Cambridge Analytica were asked to comment on the above but declined, conveniently, due to being outside of UK jurisdiction and therefore not having to follow any of our laws when it comes to privacy. This leads me to my next question regarding the Russian Ads scandal, are they really from Russia? It sounds too simple. If Russian groups really wanted to influence the US election why would they do it from a Russian IP address that’s easily traceable? But I’m digressing.

So, can we learn from Micro targeting and how it’s been used to influence voters? No. Commercially, we’re already using this strategy to identify and target consumers, then serve them individualised ads. We often collaborate with our sister agencies who help us profile customers segments that we can then serve tailored ads for. We also integrate data from Facebook’s partners such as Experian, to gain better insights about our target audience, again, allowing us to narrow targeting down. Let’s not forget, there’s going to be a lot more options in terms of targeting, when your audience, are in essence, an entire nation.

 

Just realised this is in the wrong section - apologies. Dont see a delete option?!

 

Appreciate that, we didn't care