Together with my former colleagues from Simon-Kucher led by Fabian Farkas, we just published an article in the Stanford Social Innovation Review on how behavioral economics can supercharge your online donations. Check out the article here.
Filter bubbles supercharged by social network sites and digital news platforms are widely seen as a problem. But somehow, everyone believes just “the others” are blinded by them. In this article, I will illustrate that filter bubbles may very well be humankind’s #1 challenge. The time has come to end this unintended but destructive consequence of artificial intelligence. I will show why simple regulation will not fix filter bubbles and suggest a concrete solution.
Take a moment and think back on the last ten years of your life: What has been your biggest personal learning? For me, it has been the value of compromise. I have seen too many of my tried-and-true convictions refuted or at least moderated in real life. Take for example minimum wage: Neo-classical theory told us it would only cut low-wage employment. Turns out reality is a lot more complex on the effects of minimum wage. Turns out reality is a lot more complex on a lot of things! Sorry for being a slow learner, but I finally started to understand why resilient societies are built on facilitating and sometimes enforcing compromise.
So how do we improve our ability to gain consensus? Recently, I came across a video of an experiment that really rocked my world: People in the street were asked about their opinions on a number of contested issues (like “violence used by Israel against Hamas is – or is not – morally defensible”). When showing the respondent her answers at the end of the short interview, the interviewer used a simple trick to present her the opposite answers, asking the respondent to elaborate. What do you think happened?
My guess would have been that the interviewers were beaten up in the street, but no! “A full 53% of the participants argued unequivocally for the opposite of their original attitude” (Hall, Johansson, and Strandberg, 2012).
Obviously, facts did not trigger this sudden change of heart, because facts are still subject to our very subjective interpretation: In their famous study “They Saw a Game: A Case Study, the psychologists Albert Hastorf and Hadley Cantril found that when the exact same motion picture of a college game was shown to a sample of undergraduates at each opposing school, each side perceived a different game, and their versions of the game were just as “real” as other versions were to other people.
What these experiments show is that our attitudes towards alternative viewpoints matter if we want to compromise. Good news: Those attitudes can be shaped (also shown by Leeper, Thomas, 2014). Public broadcasting (for all its deficiencies) has tried this for decades, at least to some degree, e.g., in the UK, Germany, and Japan.
Yet compromise is joining the list of endangered species these days. Polarization has been on the rise for decades, but it seems to have become a challenge of global proportions.
What is wrong with polarization you may wonder? There is evidence that a polarized environment “decreases the impact of substantive information“. In other words, facts no longer matter, party lines do. (Interestingly, scientific literacy does not inoculate against extreme viewpoints, while scientific curiosity – aka an open mind – seems to help).
Still not concerned? Some say that the laissez-faire COVID-19 response in some countries, the Brexit referendum, and of course U.S. presidential elections since 2016 have been shaped by the polarization that is fueled by “filter bubbles” on social network sites. (I am not going to withhold the potential counterargument that polarization has particularly increased in age groups that are less likely to use the internet.)
Runaway polarization risks political deadlock resulting in more global warming, more poverty, more violent fights for the proper distribution of wealth, water, and healthcare. It can also lead to more autocratic societies. It can lead to more hunger, violence, hatred, distrust, depression, and death. That’s why it is worth taking a closer look:
“Filter bubbles”(aka “echo chambers”) describe the increasing probability of
- you being only exposed to news that fit your current worldview and
- your personal news feeds becoming more and more extreme.
A fascinating, data-driven analysis by Mark Ledwich shows the traffic flows between various YouTube political channels suggested by the site itself. Apparently, social network sites have inadvertently fueled the growth of filter bubbles not only by providing an efficient means of content distribution for basically everyone (it is not without irony that you most likely read this article on a social network site) but primarily by using machine learning algorithms that dramatically exacerbate the problem. This is at the heart of this, so we need to dig deeper.
If you have not yet seen Netflix’s “The Social Dilemma” documentary, let’s take a brief look under the hood of your news feed. It’s worth spending a minute on the fundamentals of this phenomenon. Please bear with me and make an effort to understand this – it is important, really important. (Why? Because politicians around the world so far seemingly did not take the time to get it and consequently failed to act effectively!)
Naturally, digital media sites want us to keep reading, they want us to stay engaged. This is a perfectly legitimate objective for any commercial website because user engagement = time on the site = ultimately ad revenue. Since each of us responds differently to different pieces of content, they tailor each and every news feed. To do this for millions of different users, social network sites use deep learning artificial intelligence algorithms. These algorithms are constantly trained to predict the potential user engagement of every piece of content in your news feed. Training works like this: They take the content, language, and visual information of a post as input information, and then they measure actual user engagement (comments, shares, likes, etc.) as the desired outcome. Based on this closed feedback loop, the algorithms continuously predict what drives user engagement based on real-life data. This works just like Google being able to predict whether or not a picture shows a cat or a traffic light using examples that have been categorized by a human (“supervised learning”).
And this is the key reason why it is wishful thinking to assume that social network sites will fix this themselves: these algorithms are one of the cornerstones, perhaps THE key ingredient to their ongoing success!
One of the key triggers of user engagement is fake news because they travel “farther, faster, deeper, and more broadly than the truth” (as shown in the landmark study by Vosoughi, Roy, and Aral, 2018). That’s why they are prioritized by algorithms. But fake news is just one element of the problem. More importantly, extreme political views that reinforce users’ own opinions presumably follow the same path. That’s how they contribute to dangerous filter bubbles. Make no mistake: Social network sites are actively fighting fake news on various fronts, like restricting the activity of bots and adding friction to sharing certain news. But they would never abandon the core reinforcement logic that drives their news feed algorithms.
It seems like a classic “prisoners’ dilemma”: Each social network site has an overwhelming incentive to use these algorithms because everybody else does it, too.
The only way out? You guessed it. Someone has to force all of them to change. In comes government regulation.
However, in the past few years, much of the public debate and regulatory action has focused on the “fake news” aspect. For example, this year, France decided to establish a new anti-fake news agency to fight fake news coming from foreign sources (if you wonder what the 60 people initially assigned to this job can achieve, I am asking myself the same question, especially when you look at Facebook’s 10,000+ staff to fight illegal content…). Ahead of federal elections in Germany, Facebook is running an ad campaign on how they are fighting fake news:
Why the focus on fake news? Here is my little piece of conspiracy theory: Social network sites focus the discussion on fake news because that decoy is something that they can actually address. Few people seem to get that this is just a symptom of the underlying machine learning algorithms. Fixing fake news will not fix filter bubbles.
This April, the European Commission issued draft legislation on artificial intelligence and suggested “a regulatory framework for high-risk AI systems only”. However, the artificial intelligence that governs our news feeds on social network sites did not make it on the list of “prohibited” or “high-risk AI systems” (as outlined in Annex III), at least not yet. That needs to be fixed a.s.a.p. Also, the regulatory actions suggested (“requirements for high-risk Ai systems”) are very generic and focus on risk management procedures, leaving plenty of room for interpretation. If social network sites’ algorithms were to be added to the high-risk list, I would not be surprised to see this hashed out in courts for decades to come before anything happens.
We don’t have that much time anymore. We need to be much more specific when we, the citizens, address this key threat to consensus and we need to do this now.
A Counter-Algorithm for Content Display
Imagine a world…
- where digital media still give you the exciting content that you (don’t know you) want to see – but at the same time, they expose you to insights that challenge your existing beliefs in a constructive, effective manner,
- where social media fosters the effective exchange of ideas and debate by incentivizing respectful language,
- where citizens still have diverging interests, perceptions, and opinions, but are enabled to explore solutions that serve most of us.
We want to explore a solution that uses the power of machine learning instead of trying to fight or destroy it.
Science has already developed procedures for decades that effectively achieve consensus and change minds (Janis and King, 1954, recent and very relevant: Navajas, Joaquin, et al. 2019, corresponding TED Talk). Why should it not be possible to automate this and integrate it into the digital world? One challenge is that these concepts mostly rely on interpersonal contact. However, experts hypothesize that limited tweaks to algorithms may be sufficient to “limit the filter bubble effect without significantly affecting user engagement”.
Let’s summarize the scientific evidence on what we need to gain consensus:
Our starting idea is simple: To gain consensus, we need to learn to embrace the counter-arguments. But – and this is a fairly new and big “but” – research suggests that simply being exposed to counter-arguments in your news feed actually increases polarization instead of decreasing it (I routinely force myself to read articles in the Fox News app and I am living proof of that effect). This happens probably because content is mainly addressed to in-group peers. Consequently, this content tends to be extreme and insulting to dissenting opinions, because this drives engagement and group-think. However, this naturally also decreases the likelihood to convince others. As we learned from Navajas, Joaquin, et al. 2019, moderate opinions are much more likely to win over other people’s opinions.
Instead of simplistic rules and generic regulations like the one suggested by the European Commission, we suggest harnessing the predictive, self-optimizing intelligence of machine learning. This is what we think will work:
- The existing algorithms that govern the news feed stay untouched. This is necessary for any platform to remain engaging. Without these algorithms, any platform eventually becomes worthless because most content will be irrelevant for us. They fill our echo chamber with “filter bubble content”.
- Now we need to add “counter-content” that is effectively challenging our current beliefs (which are already reinforced by “filter bubble content”). How does this work? As described above, deep learning algorithms are trained to predict the engagement of any piece of content. The same algorithms can also predict whether or not a piece of “counter-content” is decreasing the likelihood of engaging with “filter bubble content”.
- The power of artificial intelligence will find persuasive tactics we may not even be aware of today. Think of it as two algorithms constantly hashing it out. Those algorithms can become much more effective than any televised U.S. presidential debate. Why? Because this algorithm will be trained not just to mobilize its own followers but also to convince other followers.
Interested in the details? Here is how AI veteran and expert Frank Buckler describes it:
- P denotes a person so that the algorithm can adapt to her interests.
- Let C be a set of information that describes a piece of content by using its text and visual information (“filter bubble content”).
- L(C, P) is the likelihood that person P will engage with content C and has to be maximized. The mathematical function that calculates L based on C and P today is shaped by social media’s deep learning algorithms. It is not necessary to understand how they work. It is important to accept that they can estimate any unknown functions that predict L based on C and P if P has interacted often enough with different kinds of content C in the past. The more the person interacts, the better the prediction becomes.
What we now suggest is to include more information:
- Let CC be a set of information that describes a second piece of content that is exposed to the person simultaneously or in close succession (“counter-content”).
- L(C | CC, P) is now the likelihood that P engages with C given CC is exposed and has to be minimized.
- The content C itself is minimizing the engagement with CC [=min L(CC | C, P)]. This makes sure that counter-content CC is contradicting and does not further exaggerate content C.
Is there a better solution?
Let’s summarize other potential solutions under discussion:
- Outlaw filter algorithms: As described above, this would impair the usefulness of content platforms so severely that this functionality is likely to happen illegally and/or indirectly. The same would happen if we outlawed filter algorithms just for politics or tried to ban political posts altogether.
- Introduce a mandatory “Driver’s License” (to use social network sites): While this may improve respectful language and help people to recognize fake news somewhat, it does not address the underlying problem: systematically misleading information and flawed learning through the selective presentation of information.
- Increase support of public broadcasting: Unless public broadcasters use similar algorithms they will never stand a chance against digital media platforms that supercharge their user engagement with deep learning algorithms.
- Mandate generic risk management for deep learning algorithms (like the proposed EU directive): Since these algorithms are mission-critical for the platforms’ success, generic legislation that leaves plenty of room for interpretation will inevitably result in decades-long court battles. Introducing a mandatory code of conduct for platforms’ use of deep-learning algorithms is likely to have the exact same effect.
This Article is Useless
…unless you comment and share it.
My intention in writing this article is to explore how we can change the world for the better. I want to directly influence policy-making on digital media. However, no article alone can achieve this. Only if readers comment and share this article, only if it becomes viral, will it have the chance ever to matter.
This is why I am asking you to comment and share your view.
This is why I am asking you to share this article as broadly as possible.
If you think this article is bogus, PLEASE COMMENT.
If you think more people should read this article, PLEASE SHARE.
If you agree with my conclusion that we need a smart solution like a counter-algorithm to save our world, PLEASE SHARE.
In any case, make up your own mind, but always remain curious.
Eating less meat is one of the top drivers to bring down our personal CO2 footprint (and no, using fewer plastic bags does not make the top ten list). That is one of the reasons why plant-based food is hyped these days. But after surveying more than 5,000 people in the US and Germany, the truth is that we massively underestimate how much meat we still consume. That begs the question of how to encourage more sustainable food habits.
If you live in the “climate crisis” filter bubble like myself, you get the impression that everyone is happily munching on plant-based food 24/7 these days (recent sales data points to exponential growth). In 2020, 400,000 more Germans claimed to be vegetarian than the year before. Yet, per capita consumption of meat across OECD countries has actually increased by close to 9% since 2010.
Now try this: When speaking with friends, casually drop the message that you have become a vegetarian. My guess is that in 8 out of 10 cases, your friends will feel compelled to respond like this: “Oh well, actually, you know, we really don’t eat meat that much anymore, do we, honey?”
It seems that being a vegetarian (or consuming less meat) is becoming the “socially desirable” answer. This is great if it heralds a new social norm of more sustainable eating behavior. However, it also means that we are more likely to lie (to others and ourselves) about our true meat consumption.
To get to the bottom of this, the Donanto Foundation sponsored a series of simple surveys in the US and Germany. Here is what we started with: “Please estimate: How much meat do you personally eat on average per day? (Please remember to include meat in frozen meals, convenience food, when eating out, cold cuts, etc.)”
We gave respondents seven answers to choose from and since people have trouble estimating meat weights, we explained each option briefly:
(Nerd note: We randomized this list of options for each respondent to eliminate potential order effects. The analysis was straightforward: If we have a sample of respondents that is representative of the overall population (aged 18+) and multiply the percentage frequency of each choice with its individual meat weight, we can add up the average per-capita meat consumption and compare it to the real average meat consumption.)
Personally, I was not surprised that respondents’ personal average was lower than the real average meat consumption. What did surprise me was just how much we are off: We actually eat 71% more meat than we think in Germany and 49% in the US.
To put this in perspective: Over the course of a year, each of us eats an additional five chicken, plus one-fifth of a pig, plus 2% of a whole cow more than we think (rough estimate based on the German meat mix). I am no psychologist, but perhaps this is an extension of the Dunning Kruger effect, where “people are typically overly optimistic when evaluating the quality of their performance on social and intellectual tasks”.
(Nerd note: You may think this is because respondents had trouble allocating their meat consumption to an average day. To check, we asked another 600 people the same question, only this time we asked for their meat consumption over an entire week. Lo and behold, the weekly respondents were even further away from reality than the daily ones!)
To test if we can eliminate the social desirability bias and tap the wisdom of the crowds, we repeated the survey with a slightly different question (in Germany): We asked respondents to estimate the daily meat consumption of the average person, not for themselves.
Well, what can I say, the wisdom of our crowd was much better than their personal estimate, but still not great overall: In reality, we still eat roughly 24% more than respondents estimated.
You probably noticed that I have been withholding the real meat consumption figures. So here comes this ugliest of truths: On average, each of us eats about 280 grams or roughly 10 oz of meat EVERY F***ING DAY! That’s in the US, in Germany, it amounts to 160 grams.
Wow. I’m going to let this sink in for a second.
Even after adjusting for some spoilage, plate waste, and other losses in grocery stores, restaurants, and homes, these numbers are still stunning, since they include everyone who cannot or will not eat meat.
OK, so excessive meat consumption is a major driver of climate change (and bad health), yet most of us fail to recognize how our own consumption is fueling this. Now let’s step back and think about what to do with this insight.
Here is a crazy thought: What about applying some learnings from smoking cessation programs? Smoking cessation has been under scientific scrutiny for decades, so there is a host of valuable knowledge out there, for example:
- Smoking cessation follows specific stages (precontemplation, contemplation, preparation, action, and maintenance), and “how much progress patients make after an intervention is directly related to what stage they are in prior to intervention”. If most meat-eaters are in denial about how much they really eat, perhaps we need to target earlier phases of their process.
- Over a hundred studies show that making tobacco more expensive is “a powerful tool for reducing tobacco use” and that tobacco taxes are not very regressive due to the high price sensitivity of low-income smokers. Therefore, the current discussion on increasing meat prices in Germany, e.g., by means of a surcharge to finance more animal-friendly farming (“Tierwohlabgabe”), may have a positive side effect.
If you are already in the “action” stage and would like to do something about your personal meat consumption, bear in mind to start small, but with consistency, for example, try to skip meat for lunch every Friday. If that works, move on to grander plans, like ProVeg’s free online 30-day Veggie Challenge. Good luck!
My dear fellow frequent travelers – COVID-19 has grounded most of us (…and deprived us of bland lounge food).
In case you are wondering what the one thing is you can do to slow the climate crisis: LET’S….JUST…DON’T…FLY…AS MUCH when lockdowns lift!
The numbers tell a stunning story (check out the graph below): The flights that I would need to retain gold status at Lufthansa cause three times the CO2 footprint of an average German citizen!
Here are three simple, evidence-based quick wins for any online fundraising effort!
Interested in the details? Check out the full deck here, with specific instructions on how to set individual donation amounts on page 5:
Some charities give potential donors lots of flexibility when donating online (see below for example 1 by Wikimedia). After all, people love flexibility, right? Then why would anyone limit donors’ choices, e.g., by pre-filling the online donation form like Unicef does it with a very structured form (see example 2 below)? Ideally, you would give donors maximum flexibility by leaving everything up to the donor, as GoFundMe does it (see example 3), right?
Well, I used to think the same way, because consumers would tell us in focus groups and interviews how much they value freedom and flexibility.
However, when we would actually test both alternatives separately (i.e., one group of consumers would only see a very flexible choice and another group of consumers would only see a very structured choice), in every single case, more consumers would choose the structured choice!
Initially, our clients and we were so shocked by these counterintuitive results that we invested in a brain scan study to find out what secret thoughts drove consumers’ choice. As we have described in an article on LinkedIn, structured choice (what we often call “bundled portfolios”) clearly outperforms the “build-it-yourself” or flexible choice in terms of mental “attention” required. “Attention” is mentally demanding and leads to choice stress, which ultimately lowers the attraction to buy.
How much lower? Well, since this issue came up in many of my projects as a management consultant for leading telecommunication and insurance firms, banks, utilities, etc. I was able to test this about ten times. The results: By deploying a flexible choice for consumers, you leave between 40% and 80% of revenue on the table as opposed to a structured choice. Those tests were conducted in the context of commercial subscription products (something I shared on LinkedIn here).
Does this apply to online donations? Initial evidence says yes: A/B pre-tests with online donations I have conducted suggest very similar outcomes!
For now, let’s note my first cardinal rule for online donations:
Do not give donors too much choice in the donation form (even if they tell you that they want that choice).
Stay tuned for more rules to come…
After I had offset the CO2 emissions of our family summer vacation, I also wanted to better understand what more we could do personally to cut our emissions. I found all sorts of resources on the web, but they seemed either much too broad or much too narrow in focus.
Instead, I was looking for a concise, quantitative, prioritized list of key CO2 emission sources that I could concentrate on. I was so intrigued by this idea I even drafted a dummy version (see screenshot below):
A number of specialized NGOs I contacted with this dummy version really liked the concept, but did not have anything ready to share. So I ended up researching the facts myself during one internal meeting that was particularly boring.
Turns out there is an incredible amount of well-researched data on CO2 emissions available out there at your fingertips! However, you really have to be a data-savvy expert to make sense of it! So I invested a few hours to research the most important facts and compiled them in a simple spreadsheet. Then I had an enthusiastic colleague (Carsten) help me out triple checking the facts and expanding the list.
That initial analysis contained quite a few surprises for me personally: For example, I had substantially under-estimated the CO2 impact of switching to a vegetarian diet! And I was very surprised to learn that avoiding plastic bags had basically zero impact on my CO2 emissions.
I started wondering if my fellow citizens had the same misconceptions on what were the key levers to cut personal CO2 emissions. Since easy-to-use insights were inaccessible and most lists lacked numbers, it seemed only natural that everyone would be just as clueless as myself.
Five surveys (with a total of more than 6,000 respondents in four countries) later, we were able to confirm that hypothesis! Our most striking finding: People across the world believe that avoiding plastic bags is actually by far the most important personal lever to cut CO2 emissions.
When we published those findings in a simple blog post on LinkedIn, things got really wild: the post generated 10x more views than my next best post and its core graph gathered thousands of likes on Reddit within a few hours. Within a few days, the story got picked up by multiple national media outlets as well as a few international ones like Wired UK or Treehugger.
The most beautiful and fun implementation of our findings was done by Handelsblatt, Germany’s leading business daily: Not only did they publish our insights in their “graph of the day” category covering two full pages, but they also created an animated version here.
Which personal action has the strongest impact on reducing the CO2 footprint of an average American or German? Turns out this question is surprisingly difficult to answer for most of us. And what are the next best five or ten things we should do to cut CO2? There is surprisingly little guidance out there. Because of that, I fear that we may not manage to cut emissions as much as we have to. That’s why Carsten and I did the legwork for you!
I have to admit that Greta got to me. Perhaps because I also have a daughter some consider to be on the spectrum as well (remember Greta’s hate speech at the UN?…“how dare you”…we get this regularly when we ask our daughter to lay the table). So what does a management consultant do if he wants to take personal action on CO2? Look for a concise, quantitative, prioritized list of key drivers that he can select from (since CO2 footprints differ by country, we looked for German information). There is a ton of information out there, but it is usually in one of the following categories:
- The “Everything-under-the-sun” list gives you too many items and does not prioritize even though the effects may differ by an order of magnitude.
- The “Energy saver” list has useful personal actions but focuses only on one aspect of your life, usually heating and/or electricity.
- The Calculator enables you to get to the bottom of your very own personal profile and simulate in great detail what effect individual actions would have. Classic German over-engineering that only a tiny group of zealots will ever use.
- The “Kill-yourself-and-your-family” list is very close to what we had in mind, but lists as the most important driver “Have one fewer child”. Seriously?! This is as factually correct as saying that if you kill yourself right now, you will achieve the biggest possible CO2 saving. Needless to say, including this lever in the infographic is a very dumb idea because it creates a backlash and destroys good intentions.
But thanks to all these lists and useful calculators, we were able to come up with the following overview quickly. It is still work-in-progress, not perfect at all, and sometimes we have conflicting data points, but hey, it’s a start.
We have grouped the actions into
- what we can start doing today — in other words, actions that don’t take too much effort, like fuel-efficient driving
- what takes more planning and preparation, like reducing the number of flights you take or switching to green electricity
- what may take considerable investments, like switching to modern heating and insulating your home better.
As a vegetarian household with top-notch heating and insulation, with an electric car, green electricity, and no daily car commute, I can check off many actions on that list. However, some drivers were new to me, like washing clothes in cold water. The good news is that if I implement all these levers, I will achieve about 75% of the reduction target defined by the government for 2030! How cool is that?!
And now comes the bad news.
We asked 1500 Americans and 1500 Germans to select from a list of seven personal actions the one that has the strongest impact on reducing the CO2 footprint of an average person. Here comes the list:
- Energy-efficient heating/cooling/insulation
- Avoid one return trip by aircraft per year
- Eat less red meat
- Fuel-efficient driving
- Buy local and seasonal produce
- Unplug unused electronics to stop standby
- No more plastic bags
We presented the list to respondents in a randomized order, but the list above is in descending order of impact. For example, in Germany, the impact of energy-efficient heating & insulation on our CO2 footprint is a whopping 250 times bigger than stopping to use plastic bags.
Guess which action was selected most often in Germany? “No more plastic bags”! Seriously?!
Here is the drama in all its gory details:
We were hoping that this is due to Germany’s unique obsession with recycling trash since the 1980s. But “no more plastic bags” actually made it to number 2 on the list in the US as well, very close behind fuel-efficient driving. Here is the comparison between both countries:
It is difficult to say who is more clueless because both countries have their specific blind spots: That “one flight less per year” comes out so low in the US is just as ridiculous as the fact that meat consumption is not seen as a source of CO2 in Germany.
I wish I had taken the survey myself before doing the math on the actual drivers. I am certain I had my fair share of ignorance! For example, I grossly overestimated the effect of having no daily commute to work.
But we are not here to poke and pry, but rather to drive action! So now that I know all the numbers, I have pledged to substantially reduce my flights!