How to fix Big Tech and Social Media

[Admiral Yi]

Your argument appears to be that the problem really isn't any problem at all. Which is a good argument against my solution, of course. Just say that the problem isn't worth the solution, and we should be fine with social media operating exactly as it does today.

For a guy who bitches about strawmen as much as you do, you sure do a lousy of job of characterizing other's positions.

I specifically said Facebook echo chambers are a problem.  Which is not quite the same thing as saying the problem isn't any problem at all.

I then asked if a subsciption model would solve this problem.  Which is not quite the same thing as saying the problem isn't worth the solution.

And none of this can be characterized as "we should be fine with social media operating exactly as it does today."

If you're not going to bother responding to what I actually post, please tell me so that I don't waste my time responding.

[Berkut]

Your argument appears to be that the problem really isn't any problem at all. Which is a good argument against my solution, of course. Just say that the problem isn't worth the solution, and we should be fine with social media operating exactly as it does today.

For a guy who bitches about strawmen as much as you do, you sure do a lousy of job of characterizing other's positions.

I apologize - I thought I was accurately characerterizing your position. It seemed to be that there wasn't a problem worthy of solving, in that you said I was overbaked or something.

I specifically said Facebook echo chambers are a problem.  Which is not quite the same thing as saying the problem isn't any problem at all.

Exactly - that was my point, which you cut out. You say there is a problem, then say there isn't a need for a solution, because...why exactly? That is what I meant by the analogy with children in coal mining. If you say its fine because nothing bad really happens, then remark about the bad things that happen, I find that somewhat confusing when I propose a solution.

I then asked if a subsciption model would solve this problem.  Which is not quite the same thing as saying the problem isn't worth the solution.

I think it would - or rather, I think it is nearly certain to be a lot better.

A subscription model would result in social media trying to appeal to the people who are using social media, rather then just trying to capture their attention so it can be sold to someone else.

If I subscribe to Facebook, you don't need me to be on facebook for another 15 minutes so you can sell that 15 minutes to someone. You just need to provide an experience for me that keeps me paying my monthly fee. If I use it 10 hours a month or 50 hours a month doesn't matter to you, so your engagement algorithms won't need to be tuned to constantly answering the question of "HOW DO WE GET BERKUT TO STAY ON ANOTHER 120 SECONDS???? AND ANOTHER??? AND ANOTHER???"

That might have its own problems, but I think it would be definitely much better then what we have now, and what we are going to have in the future.

Remember, those algorithms? They are like first generation technology right now - literally less then a decade old. And they are learning, and they get more and more and more data every single day.

This is like, I don't know, the first decade that someone invented cocaine, but where there are literally billions being invested by the very smartest people in our society in figuring out how to make better and better and better cocaine.

[Admiral Yi]

You say there is a problem, then say there isn't a need for a solution, because...why exactly?

You can't help yourself, can you?

[Berkut]

I think the issue is less about the "consumer is the product" model of advertising, and more about the fact that the most virulent propaganda can be served, that there is no control for truth, and that there is no professional standards that apply to the content that's created and promoted; that and that the AI "you liked this so try that" algorithms create effective funnels into extremist content.

The model is the problem *because* it means that serving up virulent propaganda is (due to human nature) often the best way to create more of the product. Take away the need to capture as much of the consumers attention span as possible (because that is the product) and you take away the incentive to create algorithms that exploit the worst parts of our own nature.

That butts into the free speech issue and the kind of social engineering that's dangerous - and that Americans are particularly allergic to - but that's where the actual issue lays IMO. The problem is not - as Yi rightly points out - the VW and iPhone and cereal ads particularly.

The ads are fine, actually.

I do agree there are serious free speech issues at play here, and my idea is definitely rather disturbingly "nanny state" in the worst way.

[Berkut]

You say there is a problem, then say there isn't a need for a solution, because...why exactly?

You can't help yourself, can you?

Fuck man, I am trying. I post these long responses that I try to explain why I am asking and responding the way I do, then you cut out all the explanation and leave one line and get all put out as if that is the only thing I said. If that the most interesting part of my post?

Probably better that we just don't engage. I can't do much better myself. If your goal is to score some points by figuring our how to select out a single line and respond to that, I am willing to concede that victory to you. The Yicratic method?

[Jacob]

It is a problem with social media but is it really a problem of social media?

It's always been true that people tend to associate with like-minded people and the resulting insularity can foster extremism and vulnerability to conspiracy theories.  For example, most white people living in say Mississippi in the 1850s would have held objectively extreme and conspiratorial opinions.  It was common to believe that chattel slavery was a positive good, to the benefit of the slaves and that the South was being targeted by horrific conspiracies from abolitionist zealots seeking to spread miscegenation and destroy the virtue of Southern womanhood.  Even by the standard of the times, there were extreme views and we know that led to great violence. That is perhaps an extreme example, but one can pretty easily show other examples of extreme polarization in other periods, such as the Gilded Age and the resulting civil violence.

It may be the case that the political polarization and extremism we see today is closer to the norm and that the seemingly less partisan era after WW2 was the departure from the norm.  The broad-based participation in the war fostered mixing of different kinds of people from different backgrounds and brought people together in a common pursuit.  The nature of mass media in the postwar period - eg. dominance of TV by a few corporatist networks - also worked against polarization. ( Even so, there were serious outbursts of political extremism - from McCarthyism and Jon Bircherism on the Right to the some the excesses of the New Left). 

The decentralized nature of social media helps foster insularity but it is not unique in this regard as the rise of politicized talk radio and cable news also facilitates people to choose their preferred reality.  Social media is less coherent in its messaging but it does draw some additional power from its direct peer-to-peer contact and the impact of emotive imagery ("memes" and videos).  But the broader point is even if social media disappeared overnight, the underlying problem doesn't go away.

A fair set of observations.

[Sheilbh]

I think the issue is less about the "consumer is the product" model of advertising, and more about the fact that the most virulent propaganda can be served, that there is no control for truth, and that there is no professional standards that apply to the content that's created and promoted; that and that the AI "you liked this so try that" algorithms create effective funnels into extremist content.

I don't disagree about algorithms funnelling people.

But I think there is something more to the consumer is the product angle because that ecosystem is huge and out of control. There's no transparency over really what data is really collected, how it's used or where it goes - across billions of daily transactions. When you speak to people who work in adtech it is concerning and it's an ungoverned space on the internet.

Privacy laws are beginning to impose but the existing players have such a huge advantage new entrants are difficult to imagine and it's an invisible market with minimal transparency. Google for example gets the space from publishers on exchanges, owns the exchange and sells it on the buy side to advertisers - that sort of market dynamic is problematic for all sorts of reasons. There is - rightly - minimal trust for this, so many publishers have to use Google but are basically certain that Google deliberately directs advertising revenue to their own properties like YouTube. Similarly advertisers are far from sure they're getting the best eyes on their ad (when they get human eyes at all).

It's a really fucked up, non-transparent, self-dealing part of the economy and it's a market worth billions where our data is the product. It slightly concerns me on a sort of social/global level that a few of the biggest companies in the world generally basically sell ad space. I don't think advertising has ever been that big before.

In terms of how you fix it I don't think the growing body of privacy laws are enough, I think it needs to be treated as a market on which billions of dollars are traded. So I'd suggest basically banning the owners of the exchange from selling and buying ad space on their own behalf. In an ideal world I'd also move towards a "data trust" model so the people who are able to best monetise our data are those with the best models/algorithms (including new entrants) not just the companies with the most data to begin with while also enhancing individual control over our own data.

The privacy push is leading to interesting solutions that may help - but I'm fairly dubious about what they'll actually look like.

[Josquius]

The trouble with the world today is the speed at which misinformation can spread, be updated, and come with a huge data stack to claim authenticity.

For a historic comparison see the later chapters of Homage to Catalonia. The communists are moving towards purging the anarchists. Nobody really knows what's happening. Rumours are flying all over the place of what is going on, soldiers return from months on the front line and march straight into jail their organisation having been outlawed weeks ago.
Reading this I remember that I couldn't help but wonder how it might be different in an age of social media if official channels were cut off and it was left to personal networks to disseminate and sort out rumours.

I do have faith that the digital natives and the new generation will be a lot more capable of handling social media than boomers have proven. Nonetheless we can't just wait for the boomers to die. People are suffering in the here and now so something does need to be done.

As I've said in the past I do think China had something tight with linking real ids to online ids. Some sort of government ID api needs to be set up so you can validate yourself as a real person without having to actually trust Facebook and the like with it.
Perhaps limit this to public platforms rather than one to one communication platforms. They do say whatsapp misinformation is an issue but I don't think it's quite so harmful and it must be tackled in a different and rather more traditional way.

Also seeing anti vaxers I grow increasingly keen on criminalising the deliberate spread of harmful misinformation. A delicate one for sure that would have to be set up very precisely so as not to harm free speech too much. But those arse hole doctors making a fortune from exploring the ignorant at the least need a slap.

Oh. Also micro targeting. That needs to stop. It should be outlawed. Especially when you're telling different groups incompatible things. Eg brexit and it's end to immigration /easier immigration dychotomy.

[Sheilbh]

Honestly I often just think we need a healthy dose of shame and guilt.

Maybe we went too far in the past but I feel like we could do with a modicum of shame and guilt, especially in public life

[The Minsky Moment]

But I think there is something more to the consumer is the product angle because that ecosystem is huge and out of control.

The product is information - data about the consumer and their behavior and interest.

The difficulty in designing a response stems in part of the multiple effects and implications: (1) the conflict with principles of privacy, (2) the fairness of a consumer transaction where valuable data is being acquired without the consumer understanding the nature or true value of what is being sold, (3) the competitive implications of companies that enjoy significant network effects obtaining proprietary and exclusive rights over consumer data.  Three quite different  sets of effects with different policy implications.  The pat answers to these kinds of questions that have developed over the years in the context of tangible transactions do not necessarily work the same way in the virtual context.

[Barrister]

Honestly I often just think we need a healthy dose of shame and guilt.

Maybe we went too far in the past but I feel like we could do with a modicum of shame and guilt, especially in public life

I would strenuously disagree.  If you go on to Twitter what we need are orders of magnitude less of shaming and guilting, and much more grace and charity.

[Sheilbh]

I mean guilt and shame as internally felt impulses/emotions not as something you do to others.

[Admiral Yi]

I would strenuously disagree.  If you go on to Twitter what we need are orders of magnitude less of shaming and guilting, and much more grace and charity.

I'm pretty sure Shelf meant shame and guilt in the sense of introspection and self control, not in the sense of scolding and berating others.

I do agree with him that there seems to be a correlation in increase in hectoring and decrease in self reflection.  I would go further and say the nagging of the church lady is a coping mechanism to avoid the pain of pondering the morality of your own actions.

[DGuller]

This is easier for me to address.

I don't think advertising serves a *practical* useful purpose at all. (By this I mean it doesn't serve a purpose to society in and of itself - obviously it is useful to the companies that engage in it).

Indeed, in the modern world, the only theoretical purpose it had has been supplanted anyway. That would be the purpose of informing consumers about your product, which has the purpose of giving information to consumers, which is necessary for them to make informed choices that drive the functioning free market.

The problem is that advertising is only very loosely linked to actual information. Since it is funded by businesses who by and large don't care to actually inform you, but rather to convince you, then it has always been a pretty shitty way to achieve the "inform the consumer" service anyway.

Now? I don't see how it helps me at all to have ads thrown at me all the time. I make decisions based on available customer reviews, mostly.

So yeah, I don't think advertising itself has any particular intrinsic social value that needs protecting. It's not like if you took it out of the social media model, there would not still be other avenues of advertising open.

I don't disagree that there are plenty of undesirable things that come with advertising, and the problem isn't new, but I do think that advertising serves some practical purpose.  Often times ads make you aware of solutions to problems that you never knew you had (and I don't mean that ironically).  I find that often the most difficult thing is not finding a solution to a problem, but rather identifying the problem in the first place.  Fixing a bug in your program is usually easy.  Realizing you have a bug in the first place can be very hard.

[Oexmelin]

It is a problem with social media but is it really a problem of social media?

It's always been true that people tend to associate with like-minded people and the resulting insularity can foster extremism and vulnerability to conspiracy theories.

...not really. Unless by always, you mean, only relatively recently, in an age of relatively increased mobility where certain people were entitled to voice their opinions relatively freely.

Historically, insularity tended to foster conformism, because the price to pay for dissension could be high. Small villages made up mostly of your relatives could not so easily be split by quarrel, and straight up leaving was onerous (and often life-threatening). Dissension therefore mostly expressed itself in factionalism - i.e., a confederacy of malcontents against existing order. "Bipartism" is a historical artefact of that time. Conspiracies, meanwhile, are generally considered as sociological coping mechanisms against powerlessness. Things must be coordinated by a few hidden people, because historical agency is limited to the few.

It takes gigantic, revolutionary, universalist ideas that require total committment to upend this sort of baseline. Religious reformation, liberalism, emancipation, socialism... These make demands that explicitly go beyond the small communities. Polarization isn't the natural state of communities: it's actually quite an important crisis. What democracy did was to raise the threshold of acceptable dissent. You didn't simply have the choice between silence or civil war. You could express some measure of dissent legitimately. How much dissent was the question always worked out politically.

But institutions always had a really important role to play, whether by offering a venue that structured or channelled political debate or dissent, supplying vocabulary to an ill-defined sense of "wrong", or relayed ideas. The moments of deep polarization in the US are moments of institutional crisis - the rise of abolitionism, for instance, didn't happen because anti-slavery people settled in the same area; sectionalism took root because of the institutional power and role of states in the US constitution.  Indeed, we have lots of testimonies of anti-slavery people who ended up defending slavery *after* having lived in slave societies.

There clearly is a political crisis in the US, as there has been in the past. That crisis is both fueled by, and fueling, international debate, as American crisis have done in the past - and the way out of that crisis is liable to be ugly before it is good. What's new is that there are new institutions that structure the debate, and we are woefully underprepared to deal with them as such.

Social media are no longer just corporations. They have become institutions. They structure the political debate. They relay ideas and vocabularies. Yet, we have instituted no claim on them - the way we have done so with justice, school, or even with the classical concept of "corporation". Because unlike classical corporations, social media create dynamic communities of dissent, and we haven't really had to deal with the emergence of these sorts of things for a long, long time.