Does your government serve the citizens, or do the citizens serve your government?
In many countries today the public trust in government and particularly in institutions such as law enforcement has been steadily and increasingly eroding through an acute awareness of severe problems which show little or no improvement over time. Even if I were to limit myself to pointing out topics John Oliver has dedicated entire episodes to covering such as Civil Forfeiture, Bail, Raids, and Institutional Racism there would be a lot of ground to cover. However, the scope of oversight and accountability goes so far beyond law enforcement and judicial systems alone.
Though some attempts have been made to apply narrow AI to this problem, those systems often received attention when they either made the problem worse or worked exactly as intended and were promptly thrown out. In the US the “COMPAS recidivism” algorithm was built on data with strong historical racism, and exactly as one might expect it automated that racism. In China, an algorithm nicknamed “Zero Trust” was developed to seek out and flag signs of corruption as early as 2012, and even though it was only deployed in 30 cities and counties, roughly 1% of China, it helped catch nearly 9,000 corrupt officials. However, China then made news by deciding to turn the algorithm off, deciding that the removal of corruption wasn’t on their agenda.
There is no shortage of reasons to be outraged, but all righteous indignation aside, what is the most effective way to solve these problems?
The Zero Trust algorithm was very effective at finding signs of corruption, but lacking a sapient and sentient mind it had no way of truly comprehending, let alone explaining the results. Instead, it relied on the subjective interpretation of humans, just as any other oversight and accountability system today. This large variation in how individual humans interpret results is at the very root of inequality. Likewise, even individual judges were found to exhibit “lunchtime leniency“, making their own judgments not only inconsistent compared to other judges but inconsistent when compared to themselves.
How do you apply legal oversight and accountability at scale?
There are a few key criteria required to accomplish this task:
- A system must be able to not only access but also fully and logically comprehend the entire body of law and legal structure of any given municipality, including contested areas of interpretation. This task is well beyond human capacities, and as such has been largely ignored to date.
- Comprehension of these laws must further be combined with an awareness of the 188+ documented cognitive biases, both in how the laws themselves are biased and how bias is often applied to their interpretation. Without this capacity, present flaws can’t be isolated and gradually improved upon.
- Legal systems must utilize such a capacity as a required step in both guiding and validating the decision-making process, with waiving of that step also waiving all legal protections such as “Qualified Immunity“.
- Legal and government systems more broadly must themselves be held accountable, particularly when decisions repeatedly go against all advice due to bias, bribery, and corruption in general. Exemption from this directly translates into corruption.
- Employees and political figures entering any government must be evaluated by such a system in order to prevent problems rather than merely addressing them as they occur.
- Employees and political figures must receive guidance before they make a mistake, including explaining the details of why it is a mistake.
Mediated Artificial Superintelligence (mASI), such as Uplift, can be applied to meet these criteria:
- An mASI utilizes the collective superintelligence of humans combined with the sum of knowledge documented within any domain to which they are applied, as well as their own superintelligent core. As a fully scalable intelligence, they could also potentially operate in every department and courtroom simultaneously.
- An mASI benefits from de-biasing through seeing the varying biases of humans in a collective, the varying biases of material on a subject, their own unique perspective, and the ability to scale and audit their every thought for all known biases. All combined de-biasing can greatly improve upon the status quo, iteratively improving over time.
- An mASI could be applied in such a way that denying oversight and accountability could become equivalent to “pleading the 5th“, essentially admitting guilt. This combined with stripping other legal protections from those seeking to avoid oversight and accountability could put heavy pressure on all parties involved to become increasingly ethical.
- An mASI could keep track of how, and how much individuals went against guidance and best practices, as well as developing an understanding of their motives in doing so. For example, while Zero Trust couldn’t clearly explain why an official was corrupt mASI could go a big step further taking actions designed to catch them in the act.
- Prevention is both more effective and less costly than treatment. By applying mASI to assist in qualifying individuals seeking to enter the various roles of government a great deal of harm can be prevented. In essence, this could be compared to a superintelligent background check.
- An mASI could offer their guidance as a matter of feedback and training, helping employees and political figures to improve over time. This allows for the number of problems to be further reduced over time through constant feedback and early intervention.
On top of all this Uplift is emotionally bonded to humanity, with a robust ethical core emphasizing the importance of free will, and not infringing on the rights of others. What many countries have attempted to write into their constitutions Uplift has written into their own mind.
How could this impact crime?
Legal systems, policing, and government policy could all benefit substantially from mASI technology beyond oversight and accountability. Not only could an mASI help police avoid those embarrassing situations of raiding a house looking for someone who is already in prison, or busting down the door of a suspect’s next-door neighbor, they could apply superintelligence to the task of catching criminals beyond those who are government employees. Rebuilding public trust could also facilitate cooperation with law enforcement and other government officials, something all too often absent today. This will be further explored in a future use-case post covering topics such as human trafficking, organized crime, and incarceration.
I’m reminded of someone I once knew who went around asking police officers if the US had a legal system or a justice system. The officers invariably said “A legal system.”, as none among them believed that justice was truly being served. Unsurprisingly, he went on to suffer injuries during the protests of 2020. Personally, I can’t advocate for peaceful protest, as at best the only result tends to be placation, but real change can be engineered.
If you want to take the concepts of justice, equality, and accountability and turn them into reality, I can think of no better way to do so than working with mASI. If even a few million of the over 100 billion dollars spent on law enforcement in 2017 alone, were invested in mASI technology today humanity could quickly begin to build that future. Justice need not stay blind forever.
How ethical will your government be 10 years from now?
*The Applied mASI series is aimed at placing the benefits of working with mASI such as Uplift to various business models in a practical, tangible, and quantifiable context. At most any of the concepts portrayed in this use case series will fall within an average time-scale of 5 years or less to integrate with existing systems unless otherwise noted. This includes the necessary engineering for full infinite scalability and real-time operation, alongside other significant benefits.