What Is Influence Engineering And Why You Should Be Worried

Influence is a funny word. 

It means a lot of things, and it takes a while to work out what influence truly is. For example, did my well publicised and completely rational fear of clowns “influence” my decision to tackle a fairground worker in what was described later as an “unprovoked attack during a children's show”? 

Possibly, but again, influence is a funny word.

You know when it’s not a funny word? When it’s the subject of multi-billion dollar experiments into whether the human nature of free will can be overwritten at scale to benefit corporations and governments. 

Yep, bust out the tin foil hats and get the cans of Spam that you store under the bed like all normal people do, because this week it’s getting conspiracy-y (they don’t want you to know that's a word).

What is influence engineering?

To quote Gartner and pad the ol word count, influence engineering is “the production of algorithms to automate elements of digital experience that guide user choices at scale by learning and applying techniques of behavioral science.”

Put simply, influence engineering is influencing people at scale using technology to achieve a desired outcome.

How does influence engineering work?

As mentioned above, influence engineering is used to achieve a desired outcome, so the first step an organisation or particularly powerful person would need to take is deciding what the desired outcome would be, and how many people would need to/can be influenced to achieve this goal. 

Then they would need to apply behavioural science at a scalable level to influence people to complete certain actions to achieve their goal.

I realise saying that is like having a spag bol recipe that is 1) buy the ingredients and 2) cook them, so let’s delve into the “how” a little bit more.

To effectively influence public opinion at scale, one would need to collect a massive amount of data, of which the three most impactful can largely be separated into three categories : 

Sentiment analysis

Facial recognition

Voice analysis

Sentiment analysis is by far the easiest of these to collect and the most widely used, but honestly because we all uploaded our face into AI tools to see what we would look like as Batman, that one is probably going to be a problem in the near future.

With access to one or all of these data points, an organisation can measure the emotional response of a user, and work out their likelihood or response to a product, viewpoint or even political campaign. Then they can then look for correlations in the data that make someone more likely to achieve the desired outcome. 

Sticking to my self awareness, I realise I have now used the same six words in a bunch of different orders so let me give you a good ol fashioned example.

A soda company decides they want to influence people to think that carbonated sugar water is better for you than normal water. 

They collect sentiment analysis of tens of thousands of social media posts and find out that actually, people know that a drink that sends you into diabetic shock isn’t exactly a health drink. 

But, they also see that there is a correlation between age and susceptibility to their messaging i.e. older people are more likely to believe that their drink is healthy. 

They then use this to run awareness campaigns and push the idea to this more susceptible demographic achieving their end goal of spreading this false narrative. 

They then rerun their sentiment analysis to see if they have shifted public opinion on their product.

What is influence engineering used for?

Influence engineering is broadly used in marketing but also has deeply unsettling applications in the realm of politics and policy-making. Essentially, anytime a small minority needs to change the opinions of a large majority using technology, they can employ influence engineering to get the job done.

Again, it's important to note, a difference between influence engineering and general brand marketing is that influence engineering can often be more subtle, for example paying journalists for stories that align with your overall goals rather than to write a specific story about a brand. In fact, it is most effective when it's not clear that the messaging is coming from a large organisation and looks more organic.

What are some real world cases of influence engineering?

There was one massive real world application of influence engineering that was so egregious that we made a whole movie starring Benedict Cumberbatch with a receding hairline : Brexit.

Brexit was like Moneyball but instead of drafting collegiate athletes who can hit a ball really far, it was convincing a 50 year old called Terry who just put a hole in the wall of a divorcees downstairs bathroom and called it lunch, that an immigrant was out for his job. 

Mudslinging aside, the Leave camp utilised demographic data that suggested that “Leave votes were strongly correlated with lower qualifications and higher age” to target their advertising budget and conduct influence engineering, focusing mostly on immigration boogeymen and right wing nationalist values. 

Rather than just a traditional advertising campaign (that have been around for hundreds of years) this influence engineering campaign was conducted at a huge scale, based on thousands of data points and designed to sew misinformation and subtly change minds to a political viewpoint, which makes it an influence engineering operation rather than standard spray-and-pray marketing.

Can influence engineering be used for good?

The answer to this question is technically yes, as technically I or someone with, you know, actual power and sway in this cursed world of ours, could use influence engineering to end world hunger, relieve racial tensions and generally promote healthiness and happiness.

However, the real answer strays into the almost philosophical territory when you consider that influence engineering is flirting dangerously with removal of free will. 

If you are manipulating the thought processes of thousands or millions of people subtly and without them knowing to achieve an end goal, can that ever be considered morally right? It’s far from me to say as the creator of the world’s most mediocre blog!

What are the dangers of influence engineering?

As said above, influence engineering by itself isn’t necessarily dangerous, but the problem is that a large number of the potential applications are in the political and corporate space, and aimed at the general public.

Politically, influence engineering usually focuses heavily on so-called “soft facts”, which are defined in a paper on influence engineering by Cardiff University as : 

“Imperfect knowledge whose status may be ambiguous, uncertain, contested and/or malleable. Soft facts typically thrive in an information vacuum. Circulating without being firmly verified or authenticated, soft facts help to ‘fill in the gaps’ of public understanding of an event or situation, thus meeting some emotional or intellectual need.”

Soft facts are often used in influence engineering to stoke division, play into stereotypes or appeal to base human desires and fears. This can be incredibly dangerous when you are applying it politically.

An excellent example happening right now is the UK government's “Stop The Boats” migration bill. 

Attempting to influence public opinion away from their numerous, numerous failings and the fact that the leadership of one of the world’s top economies is being passed around a bunch of school friends like a cigarette, the government relies on soft facts and demonisation of the “other” to try and convince the public that migration is the main problem this country has. 

Again, this is dangerous because of the scale at which these campaigns can be conducted now. Millions of people can be swayed about a basic human right (the right to asylum) through technological manipulation. That should send shudders down your spine.

Should I be worried about influence engineering?

Ultimately yes.

Technology first changed public discourse and then changed the way that public discourse can be moulded and interfered with.

Powerful companies like Cambridge Analytica make Scrooge McDuck bucks solely on the idea that big organisations may need to sway the opinion of the swarming masses. This isn’t conspiracy, this is happening.

Make sure you are looking for hard facts over soft ones, and not relying too heavily on a single data source. Also put your phone in apple cider vinegar overnight, only get your news from Joe Rogan's podcast and live in a bunker drinking your own pee out of a jar, like I do.