The algorithm, the asshole, and the virus

Social media is killing us with COVID-19

In December 2017, six months after writing a paper about how islands like New Zealand should use complete border closure as a rational pandemic response, I was giving a talk about artificial intelligence as a threat to democracy and free will to a room full of philosophers in Dunedin, New Zealand.

At the time the world had never heard of a company called Cambridge Analytica nor had any inkling of SARS-CoV-2. Unknown to me, at that exact time, Jaron Lanier was writing a (much better than my talk!) book about the malignant impact of social media on our wellbeing and society. It was called ‘10 Arguments for Deleting Your Social Media Accounts Right Now’ (Note, Lanier is not the titular asshole of this story).

After dabbling in the philosophy of AI, I returned to pandemic threats, publishing a blog in November 2019. Right when a bat was infecting the first Wuhan citizen with SARS-CoV-2, the blog provided commentary on the Global Health Security Index (GHSI), an index that conveyed a grim assessment of pandemic preparedness around the world.

In the present blog, I want to draw these disparate and seemingly unrelated related strands together. My thesis is that the business model of social media has played a critical causal role in the deaths of probably hundreds of thousands of people due to COVID-19.

‘Delete your social media right now’

Lanier was an early virtual reality developer, and has been involved with Internet2, Google (which bought his company), Second Life, LinkedIn, and a host of other digital projects. He is also a classical composer. He argued in his 2018 book that we would all be better off deleting all our social media accounts ‘right now’ and he has no social media himself at all.

In his entertaining yet serious way Lanier describes the business model of many social media platforms as: ‘Behaviors of Users Modified, and Made into an Empire for Rent’, aka BUMMER.

As we have come to understand, the social media platforms effectively sell users attention to advertisers (Lanier says ‘manipulators’) and the platforms deploy ever evolving algorithms that serve up content shown to enhance ‘engagement’.

The algorithm

What was not foreseen, but is now well understood, is that the algorithms soon learned that serving up individually tailored, controversial, emotional and negative content not only enhanced engagement, but turned us all into self-obsessed assholes, at the same time undermining truth, empathy and happiness (among other things). The details of all this can be found in Lanier’s book. Furthermore, it was not only this Unforeseen Disaster that has led to these problems, we also see Gaming the System on the part of content creators (eg Media hacks) and subversive elements (eg Russian trolls)

Lanier’s arguments for deleting all social media boil down to the following list of negative effects that the BUMMER business model has on individuals and society.

  1. We’re losing free will
  2. We must resist the insanity of our times
  3. We’re becoming assholes
  4. We’re losing truth
  5. What we say is becoming meaningless
  6. We’re losing our empathy
  7. We’re becoming unhappy
  8. We’re losing economic dignity
  9. Politics is becoming impossible
  10. We’re losing our special personhood

One example Lanier gives is the emergence of the Black Lives Matter movement (the first time around), which was facilitated by the internet and social media. So far so good. But then the algorithms determined that there was a subgroup of white, right-leaning, American nationalists who engaged tremendously with the platform whenever they were served BLM content. No doubt this is how future civil wars begin.

Lanier even gives the example of social media’s causal role in the malignant growth of the antivax movement. A perverse effect of the BUMMER algorithms is not only do they serve up antivax material because it pushes the right (or wrong!) buttons for various people, but then online marketplaces like Amazon will serve up antivax book suggestions because the user has been reading antivax material! This digital perpetual engine drives resoundingly crackpot content onto the bestseller list. Lanier puts it plainly in ’10 Arguments’ when he states that ‘BUMMER kills’. It literally does.

Which brings us to COVID-19.

The asshole

Donald Trump (the asshole of this story), having evolved into even more of an asshole through his literal addiction to Twitter and its empathy destroying asshole-o-genic psychological impacts, proceeded to divide the United States (and therefore the world) on almost every issue to do with COVID-19. Amplification of this messaging by social media algorithms, which served each item to those most likely to be pissed off by it, and by media outlets who created content most likely to piss everyone off (they know how they algorithms work and want their hits), consolidated the in-group/out-group psychology of left and right. Suddenly, a million people are dead worldwide in part because angry Trumpers (or Bolsonaros, or ‘Sovereign Individual’ Australians) won’t wear masks or stay home (nor sacrifice anything of relatively minor import in the interests of public health).

Remember the GHSI, I mentioned at the outset, used for scoring health security? Well the USA topped the index with 83.5/100. Ironic, I know.

The virus

In an article published by Sawyer Crosby (and others) on the ‘Think Global Health’ website, the authors ridicule the GHSI. They basically argue that we clearly have no understanding of how to measure health security if there is demonstrably no correlation between GHSI scores and COVID-19 outcomes (in fact, as at 31 July 2020, there was a correlation, and the GHSI was positively correlated with COVID-19 death per capita – at first glance it literally couldn’t get much worse for valid measurement).

However, any attempt to rank COVID-19 responses is surely premature. In the first instance, we don’t actually know how many cases there are globally (we do know there are at least 5 million in the USA!) An MIT study indicates that there are probably more than 100 million cases worldwide. This means that present estimates such as those compiled by worldometer are out by an order of magnitude due to undercount. Many countries are simply unable to count their cases as noted by this article in the Guardian. The flip side of a low GHSI score is that the country most likely has a low capacity for situation awareness as the pandemic hits (hence reports ‘low cases’).

However, I reiterate, we know for sure that nearly 200,000 US citizens have died. Other reasons why we can’t yet know which countries are doing well include:

  • COVID-19 ‘success’ will depend on the strategy chosen by each country (eg, exclusion, elimination, suppression, or mitigation)
  • countries impacted later can learn from those impacted earlier
  • the pandemic is still accelerating
  • countries that have done well so far may yet be overwhelmed
  • countries performing ‘poorly’ at one point in time may yet look successful in the future, eg, if they develop vaccines and roll out vaccination quickly

So, do we know how to measure (and therefore construct) global health security? Yes. But building the boat and sewing the sails and taming the wind will not get us anywhere if social media is telling people to jump overboard.

Let me be clear. No one invented social media with the intent of throwing people overboard, but the interaction of the algorithms, the assholes and the virus have almost certainly amplified the number of deaths in the United States and likely elsewhere. And those with an interest in a weakened USA are likely fanning the flames. We don’t know yet if the literal attacks on public health professionals in the US, as described in the Journal of the American Medical Association, were incited by domestic division, or Russian bot armies.

The USA should have been the country best positioned to deal with COVID-19 effectively and safely. Yet social media platforms have hijacked our cognitive biases, produced a general decline in civility in political discourse, and hardened value conflicts. Look at this figure which contrasts a simulation where a group of people with disparate opinions interact (on the left) to form an agreed opinion, with a simulation (on the right) where algorithmic bias feeds content similar to already held beliefs to the individuals. The result in the presence of algorithmic bias, is a persisting dichotomy of opinion (study available here).

We know that people follow public health guidance when they believe that officials understand the public’s values and that ‘people like me’ can help make decisions. When everyone scrolls through a highly individualised social media feed, there is no such thing as ‘people like me’.

In attempting to attain health security we have not adequately accounted for social media and the machinations of BUMMER. We did not consider that people might drink bleach because the president suggested so. We have tried to include measures of ‘political stability’ in the calculus of health security, but we did not consider the supreme assholery of ‘sovereign individuals’.

The future

In the big picture, COVID-19 is minor blip. The world will go on. But if there ever was a truly existential threat, a devastating bioweapon unleashed across the globe that required everyone to take action to ensure planetary safety, then under the current regimen we are screwed. We need a mechanism for coordinating or it will be the end of us. BUMMER must end. Health security must demand it. We need a period of wisdom and coordination to prepare for the greatest of threats. The alternative is that social media explains the Fermi Paradox.

Policy Quarterly Journal: Focus on COVID-19 and New Zealand

The Victoria University of Wellington policy journal Policy Quarterly has just published a theme issue examining COVID-19 and the New Zealand context.

Articles cover: government, law, globalisation, health, lockdown, the economy and much more.

Included is a paper on ‘Public Health Aspects of the Covid-19 Response and Opportunities for the Post-Pandemic Era

In this paper (which I co-author), we argue that New Zealand’s health outcome appears to be the best in the OECD, but that some other countries made better use of certain control strategies. For example, Taiwan avoided an intense lockdown by focussing more strongly on immediate and intensive border control measures. Taiwan also excelled at using digital contact tracing and widespread use of face masks.

We suggest that in the post-pandemic era there are many emerging opportunities for society to be gained by embedding better plans for controlling future emerging diseases, and strengthening public health infrastructure.

Moving forward there is an opportunity to embed other important changes such as a ‘green reset’ and other pro-equity health interventions in system changes that will naturally follow the COVID-19 pandemic.