The internet amnesty: A proposal

It is the distant future, the year is 2028. You can pay to have an AI trawl through the internet- even the stuff that’s been deleted is recorded- to find evidence of wrongdoing by a target. It can figure out people’s pseudonyms and online identities by matching writing, track down stuff you said when you were 17, read all your tweets, etc. etc. Would it find any dirt on you? Would it find any dirt on me? Although I can honestly say I can’t think of what it would find, I’m pretty sure it would find something, I’ve been hanging out here for a long time. I’m pretty sure the Scandal-O-Matic will exist, and relatively soon, but whether it does or it doesn’t, we’ve never been better at trashing reputations than we are now, and we’re going to get better.

The root of the problem is simple enough. Between all the many different ways a person can fuck up I’d guess almost everyone in public life has committed at least one major oopsie. Now oopsie may be a cute word, but in a lot of cases, what we’re talking about ain’t cute. In the internet age, those oopsies blow up with greater and greater frequency, and as alluded to, the technology for oopsie hunting will improve.

When someone’s prior moral errors leak, it gives us negative information about their character. Everyone has made some mistakes, and most of us have made some quite bad ones, so the existence of a mistake is not necessarily a big deal in and of itself. However, In the past, when one of those serious mistakes leaked, it was relatively good evidence that someone was a bad person because you usually had to commit a fair few sins before one gets out. Now that everything is leakier, the evidence a mistake coming to light represents is weaker. However, our intuitions still haven’t adapted to that change, if they ever will.

How should we act and think in this reputation-melee? Option one is to accept this is going to be the way things are now. It may be harsh and a bit random whether any given person gets caught, but since scumbags commit more oopsies than non-scumbags, the reputation-melee will at least flush out more scumbags than upstanding members of society.

There are a few problems with this logic:

1. Scumbag reputational resilience. There seems to be a pattern whereby people who are obviously scumbags survive the discovery of evidence proving their scumbaggery.

2. Blackmail. The status quo makes people very vulnerable to blackmail. I suspect we will see more of this in the future.

3. Evil causes are particularly good at hunting down evidence of wrongdoing and using it against their enemies. In general, evil people are and are likely to dig up the wrongs of their enemies as weapons and use them effectively.

4. Scandal-focused discourse to favor cruel people. In an environment where there’s a tonne of potential scandals, those who take glee in hunting for scandals- mostly bad people- will flourish. Additionally, even if it didn’t favor malicious gossip, scandal-focused discourse is unproductive.

5. Scumbags are better at covering up their scumbaggery than ordinary people. Bad people are often better at systematically covering up the evidence than good people, which tends to neutralize the advantage good people get in having left less evidence around the place that they’re guilty.

6. The slander-media ecosystem is just going to keep getting more and more intense, are you really comfortable with that?

Finally, although not everyone will agree with me on this, I think it’s many times worse for a fundamentally good person to have a bad reputation than for a fundamentally bad person to have a good reputation. People sometimes scoff at this “what does it matter if people call you names on the internet?” To this I would reply are you human or beast? Caring about your reputation is fundamentally part of what it means to be human, even in cases where ‘rationally’ it shouldn’t matter. I would prefer to lose an arm than be publicly reviled. There is evidence that ostracism and fear of ostracism are among the most psychologically painful experiences there are.

My counter-proposal, option two, is that we declare a blanket amnesty for everything unless it’s abominable. Somewhat creepy behavior plausibly the result of misjudgment? Amnesty. Rape someone? No amnesty. Do a dodgy paraphrase for convenience. Amnesty. Steal a manuscript from another scholar and publish it under your name? No amnesty. Improperly make expense claims? Amnesty. Embezzle millions from your not-for-profit? No amnesty. My general position is that in the internet age, you should set a very high bar of wrongdoing, and not pursue anything that falls under that. I’ll admit, I’ve sometimes been a hypocrite in this matter, but I’ll try to do better. [I’d also add that, even in the no-amnesty cases, the demand should be that the person leaves public life, not that they be hounded even in seculsion].

Things aren’t as simple as I’m making out. Between my amnesty and no amnesty examples, there is quite a gap. Exactly where to draw the line is a hard question (although I’d say the line should be closer to the no-amnesty examples than the amnesty examples). Also, there’s a real danger in giving unlimited licenses for moderate-level wrongs, so at a certain quantity, we must start to treat a bundle of moderate wrongs as an abominable wrong, although I’d urge setting this bar very high, mostly because of the feeding frenzy dynamics that hunting for moderate wrongs can lead to.

What I can say is that I’m not sure exactly where the new line should lie, I’m confident that it is further in the direction of my proposal than current practice.

Edit- clarifications:

1. Some commentators have noted that what is abominable will vary from person to person. This is correct and unavoidable, thus “don’t bring up past errors unless they are abominable” can only be a rule of thumb, not a complete focal point for a new norm. I was aware of this point but should have been clearer about it.

2. That something is abominable is a necessary condition for it to be used against someone, but it may not be a sufficient condition. If X was abominable but happened a long time ago, or was only borderline abominable, or happened in a context of mitigating circumstances, or most likely, some combination of these, then it may be improper to bring up X.