Loosely applying an infamous thought experiment to the Internet
Warning! — Risk of Future Blackmail Ahead
Back in 2010, the user Roko on the philosophy-oriented blog site LessWrong posted a thought experiment concerning an all-powerful AI construct. To briefly summarize, this AI would be incentivized to devise a means to reach backwards through time, likely via simulation, and punish those who did not act to bring about its existence. The ‘basilisk’ part refers to the idea that simply being aware of this theoretical AI was sufficient to hold one accountable for their action (or inaction).
Shortly after the thought experiment was posted, some mild controversy was stirred in response to the information hazard it posed and the post was removed. Additionally, there were several criticisms levied against the thought experiment, chiefly concerning the practicality of said AI devoting resources to retroactive punishment. Despite the theory’s constraints, Roko’s Basilisk illustrates a concept that deserves further exploration.
In order to avoid, or at least minimize the risk of, misrepresenting the concepts surrounding the Basilisk, such as acausal trade [threats], I will describe the theme which serves as the focal point of this piece under the general term “future blackmail”. This theme can be drawn from the threat of future punishment for acting or failing to act on information learned at some point in the past. In Roko’s example, if you read in a magazine (or Medium article) about how an all-powerful AI might one day be reality, and you decide to not devote your life’s work to helping create it, despite knowing there’s a change it will punish you for refusing to assist, you are now in grave danger. At least in theory.
Now, perhaps you took such threats of violence seriously and changed the trajectory of your life from car salesman to AI researcher. Would that be considered blackmail? If so, could you not call it future blackmail? This is the essential theme being examined. How theoretical future threats and risks can alter one’s actions in a manner separate from traditional risk avoidance.
Risk Avoidance vs. Future Blackmail
So what is special about future blackmail? At a glance, it appears to be normal risk avoidance, a practice all of us partake in to varying degrees every day. Whether it is looking both ways before crossing the street or diversifying your investment portfolio, the desire to mitigate risk exposure modulates our behavior to avoid an unpleasant future. So how is future blackmail unique?
The uniqueness of future blackmail is derived from the practically conscious nature of its enforcer. By this I refer to the purposeful design through which punishment is exacted against transgressors. This takes the form of the Basilisk within the confines of the thought experiment, but takes another in every day life, i.e. the Internet itself and its more social environments in particular.
Herein lies the central position: that Roko’s Basilisk already exists and the Internet embodies both the creature itself and the mechanisms by which it exacts its punishment.
Our Own Basilisk
Once something is on the Internet, it never disappears.
So how does the Internet, and namely social media, map to Roko’s Basilisk? One must concede three basic tenets concerning the Internet:
- The Internet is a permanent repository
- The Internet is the universal agora
- The Internet influences the non-digital sphere
While these three tenets are far from original and are self-evident enough to border on truism, they are the essential ingredients in creating the Basilisk. First, the permanent nature of the Internet means the Basilisk has a perfect memory. Second, the universal agora ensures that dissent is inevitable. And finally, the Internet is inseparable from modern culture, borderline inescapable.
Each cascades into the other, with infinite opportunities arises for a given piece of content causing dissent in the agora, which in turn can impact life outside of the Internet. Perhaps the most poignant example of this is what is colloquially called “Cancel Culture”. For the sake of ease, this term will be used for the same pattern detection, dissension, and deletion of the victim’s of the modern Basilisk.
A Social Creature
The history of this Basilisk’s birth and development is complex and driven by a myriad of different psychologies and social mechanisms, such as the Overton Window, that fall outside the scope and intent of this piece. What is of particular interest for this cyber Social Basilisk is the potential effect it has on those who can simultaneously be considered its victims and its caretakers. The fear of being “cancelled” has had a measurable effect, with many suffering its wrath for failing to act in the past in accordance to what is deemed conducive to its existence now.
Even more interesting is the untold numbers who have taken proactive measures to avoid future wrath of the Social Basilisk, in effect future blackmail. Building on this is the phenomena of those seemingly impervious to the wrath of the Basilisk. In the original experiment, it was pointed out that those who simply chose not to care about Roko’s AI Basilisk would be invulnerable to the simulated punishment it would endeavor to wreak. This is also seen in regards to the Social Basilisk, with would-be victims of cancel culture simply ignoring it and allowing time to settle the fervor. By outward appearance, the pattern holds true.
The future of the Basilisk is uncertain, with examples indicating the creature is self-terminating, eating itself as if by nature. As such, it is not the purpose of this piece to assign morality to the Basilisk, rather to describe the peculiarity of the creature itself as well as note that the dangers of such an entity need not lie in some advanced AI in the far flung future, rather we are capable of creating our own.