I’m just trying to take the ideas in this thread seriously. If we are going to argue that it’s bad to try to influence people’s behavior, we ought to think through what that means when applied to entities other than Facebook. That will help us figure out if we really mean what we are saying. Is it an actual value we hold, or only a rhetorical point when we are taking about a few tech companies?
No it is fine sending a comment expressing opinion. There are so many qualitatives differences:
- This is human to human communication, not an optimised algorithm.
- People here are not sending you 10 comments and looking at the one which you engaged the most time, because they don't have the metrics. Social media knows how many seconds you looked at a post and classify it with AI, they can make a profile of your current sentiment.
- The poster probably didn't analyse your whole past discussion history with algorithms before answering in a way that would be best calibrated for your way of thinking.
Interactions, discussions between people, even conflictual (not violent), are fine. Communication is fine. Mass communication is mostly fine (traditional advertising, broadcast...). What's not fine is manipulating people by mass-communicating while adjusting the message given metadata that those people are not even aware about.
In a conversation, you subconsciously build a model of the mind of person you are talking to, and your words are tuned to fill the gap between what they know and are thinking and the ideas you are trying to express. If I think back through my life, the communication that has affected me most as been from humans in face-to-face conversations, not on websites. But I could be bad at objectively evaluating this.
I think for Facebook to be more effective than humans at influencing behavior, they have to implement theory of mind, which might require AGI.
I also think a human can be much better at influencing people 1-on-1. But adapting his speech to two people at the same time? That's more difficult. There's only so many tricks you can use to have two different people hear what they want to hear. 5 people? 10? 2 billions?
That's the scary thing. Facebook is good enough at influencing someone, but they do it for half the population of the world simultaneously.
In a conversation, the influence goes both ways: if you try to persuade me of a political point of view, I can do a number of things: (and whether these are valid or not, they will impact the flow of argument, and the success of the persuasion)
- I can suggest you're biased
- I can try to change the subject
- I can decide you're a bad person for holding the "wrong" views
- I can directly argue against your point
- I can attack the structure of your argument
- I can make an emotional appeal regarding why I must hold my point of view
- etc ....
In these and other ways, I can push back and modify how much I'm influenced by a one-to-one argument. I won't always be successful, for sure. But sometimes as well the persuasion go will go the other way, and I will influence you.
This is not true on social media because another power is dictating which one-to-one conversations may happen in the first place, and then loading the deck with idea before those one-to-one conversations even happen. Further, social media changes the scale of communications. If we worked together, and saw each other regularly, we could mediate each other's influence, place boundaries, etc. With social media, there is always a crowd of strangers: too many people converse with, know, and set boundaries with.
There are obviously other distinctions between one-to-one conversation and social media, but I particularly wanted to talk about the key differences here: lack of real back and forth, and scale.
Look at the documentary... One of the guy explains that AI is not yet good enough to surplace humans at their strengths, but can game us on our weaknesses. This is the exact point that people think it's not having an impact on them, while it has, even so subtle.
As a person, you can get better at convincing people of course, and masses of people even - humanity has gone through that with ups and downs. Now we have a system that's working totally differently. From example, I remind reading that many conspiracists build distance with their friends and family as they close themselves to in-person discussion or any argumentation not fitting their views.
IMHO people are too complicated, and situations too situation specific to be able to generalize about how someone is influenced.
Sometimes people ignore the advice of those close to them, sometimes they don't, and, it probably depends on the subject being discussed, the particular relationship between the people, the state of mind of the person in question, and so on.
A single person generally has orders of magnitude less power to influence than the the biggest companies in the world. That's surely not a fair comparison.
Intent matters. Intent of 1:1 conversations between people is usually win-win, or at least win-neutral. Advertising beyond the point of informing that a product exists is exploitative; it's intent is very much win-lose.
Put another way: if your friend came to you and started manipulating and pressuring you the way ads do, for the reason ads do, they'd very quickly stop being your friend.
I don't think that quiet puts the finger on the distinction.
Apple trying to sell me an iPhone claiming it's more secure and more privacy respecting might be true so if buy one it's a win-win if (a) it is true and (b) I actually wanted those features.
> Advertising beyond the point of informing that a product exists is exploitative
Okay, so, Apple showing silhouettes dancing was exploitative? The should only say "we made a device, it plays music, it's this size, the batteries last this long"?
Was the 1984 ad exploitive? I'm just trying to think of famous ads. Is the Ikea add "Time To Leave Home" ad exploitive?
> Apple trying to sell me an iPhone claiming it's more secure and more privacy respecting might be true so if buy one it's a win-win if (a) it is true and (b) I actually wanted those features.
Sure, it's a win-win if the features match your needs. And a win-win would be Apple making the claim, listing the privacy features of their new iPhone, along with honest arguments why these features protect your privacy better than competition. I absolutely do not mind things like that - it's the socially-useful purpose of advertising: informing people about products and their features, so that individuals can pick the best solution to their problems.
It's only when advertising tries to override individual's agency when I consider them bad. And note that purposefully and covertly overriding someone's decision making capability is a malicious behavior, and very rarely justifiable.
> The should only say "we made a device, it plays music, it's this size, the batteries last this long"?
Would it be bad if they did only that?
> Was the 1984 ad exploitive?
Of course. While it's nothing compared to today's ads, it still tried to sell you a computer by tricking your mind with completely irrelevant references to 1984 and the feelings it evoked. It tried to override your capability for thinking, by sneaking in an emotional payload.
> Is the Ikea add "Time To Leave Home" ad exploitive?
It's a fun comedy sketch, but again: it tries to use an emotional payload to get you into the market for furniture and think of IKEA in particular (and make it a first association in your mind, over competitors).
Look at it from this point of view: imagine you live in a small town, and there are two small-time shops building and selling farming widgets. Would you want them to spend all their energy and innovation capacity one-upping each other in comedy stand-ups on the storefront, or would you prefer them to focus on designing better farming widgets?
Context and details matter. As does intent. Saying that all discussion is really "influencing", isn't useful, imho.
In the context of our discussion about Facebook, IMHO these are the defining qualities specific to Facebook:
1. Massive, global scale and scope - behavior via Facebook can influence culture, politics, etc.
2. An algorithm driving the bus, as opposed to people (albeit people choose what the algorithm does or doesn't do).
3. Facebook's primary motivation in influencing is to profit via increased user engagement - they currently have no financial incentive to care about the particular nature of said engagement.
4. Facebook's poor track record when it comes to acknowledging critique or concerns around the power its platform has, and Facebook's denial that it should exercise a degree of responsiblity (or be legally held accountable in some way) for said power.
5. The somewhat covert nature of how Facebook functions; as this documentary shows, while how Facebook functions may have been "in the open", it's not something most non-tech folk are aware of, and it's not something Facebook has been, in good faith, forthright and transparent about.
6. The particular power and impact of computers/the internet as a medium of mass communication, which we are still learning about.
6.5 As a subsection of item 6, the viral nature of the internet/social media which means stuff spreads very quickly, unlike other modes or media of human communication.
So sure, a parent "influences" their child, a teacher influences a student, newspapers influence people, and so do Coca-Cola commercials, but not the same way that Facebook has "influence", which I've attempted to describe above.
