If the deck is stacked, change the rules
In the last few posts we’ve talked about the underlying drivers of fake news and why it is inexorably linked to the publisher economics problem.
This time let’s focus on how we can solve these problems. First, a quick recap:
- Fake news is decentralised. It is a completely disaggregated problem-set, with fake news creators operating across boundaries, with no barriers to entry, and with numerous product and distribution platforms.
- Fake news is hard to define. It is a deeply amorphous problem where errors of omission, genuine errors of commission, entirely justified albeit inflammatory opinion are intermingled with ‘true fake news’, i.e., stories published with the specific aim to mislead.
- Fake news is highly disruptive. The economics of free ad-funded news unfairly benefit fake news providers more than they do real news providers. In this, fake news is like any other disruptive technology. And like any other disruptor it cannot be defeated through ‘competition’ because the disruptor is playing by a different set of rules.
So, how do we solve fake news? In order to answer this question we can evaluate four different approaches to problem-solving. Let’s call these four approaches the Solution Stack:
L1: PERSONNEL — The top-most layer of the stack is where we solve problems through the application of human capital (i.e., hiring and performance management).
The Personnel layer is where most of us spend our waking and working hours. We hire staff, manage them, and build cultures to achieve our stated goals. This approach is relatively simple (at least in theory), quick, and cheap to implement. In the short-term.
However, a personnel-based approach cannot be used to solve fake news because the problem we’re tackling is utterly decentralised. It exists in every corner of the world, and extends to every domain of human knowledge. A veritable army would be required to fact-check and correct the record on every website, and even then it’s questionable whether the army would succeed.
L2: PRODUCT — The second layer of the stack is where problems are solved by creating novel tools (i.e., innovation and product development).
Product-based solutions require more time, financial investment and iteration to get right. But they deliver superior returns over the longer term through automation, accuracy, repeatability, and economies of scale.
The problem with fake news is that economies of scale are actually counterproductive. Not only is this problem decentralised, but all product-based solutions run counter to the commercial interests of the companies that would be required to implement them. Fake news is a highly effective delivery device for ad impressions and engagement device for social networks. So cutting it out will reduce the available ad inventory, content velocity, and emotional engagement that all of these products are measured by.
L3: PROTOCOL — The third layer of the stack is where groups agree on a set of behaviours to solve a shared problem (e.g., Shipping Containers)
Protocol-based solutions are built through consensus, which can make them hard to launch. However, they are better suited to distributed problems that can’t be solved at the Personnel or Product layers. Protocols, once established, can also be extremely durable and resistant to disruption. Take for example the fact that we still shake hands when we meet — a protocol begun in ancient Greece, and believed to be a signal that neither person was carrying a weapon.
Protocols are where we see the greatest likelihood of a sustainable solution for fake news. I’ll return to this again at the end, and in my next post.
L4: POLICY — The bottom-most layer of the stack is where problems are solved by establishing policies or regulations that everyone HAS TO abide by (e.g., financial regulation, GDPR).
Policy-based solutions can be quicker to implement because they require less broad consensus. They are also more strictly adhered to because of punitive measures. However, these solutions are also very rigid.
In the case of fake news, where even the definition of the term itself engenders disagreement, a rigid or binary solution is unsuitable. Moreover, the public interest test that is generally applicable to policy changes is particularly difficult to administer in the highly amorphous fake news context.
At present, there are numerous efforts underway around the world to solve fake news through Personnel, Product and Policy initiatives. For the reasons outlined above these efforts are highly unlikely to succeed.
A better alternative is to design a protocol, a shared set of incentive-led behaviours, that encourages the sorts of actions required to eliminate this problem.
In the next (and final) post of this series I will share the design for this protocol. It’s something that Nic Hodges, Rohin Aggarwal and I have been working on for the better part of the past year. It is, we believe, the best and most promising way to solve what has emerged as one of the biggest problems in the world today.
If you’ve already read all the posts in this series, thank you. And please do write back with your thoughts and feedback on what you’ve read. If you’ve missed the 3 earlier posts and want to catch up, here are the links: