In the second episode of the “Human 2040” series, “I Organise Myself”, we look at socio-political issues, including how elections might play out, how we will build our neighborhoods and what the judicial system might look like. In order to create a reliable picture of these aspects of life in 2040, the analysts from Polityka Insight took into account trends such as the growing role of deliberative democracy, the development of information bubbles and the phenomenon of infocation (i.e. the problem of information overload, including difficulties with focusing on important data), or the increased importance of artificial intelligence in many aspects of socio-political life. Andrzej Bobiński, managing director of Polityka Insight, talks about whether the presented vision of organizing ourselves in 2040 has a chance to come to fruition in the first podcast with Professor Aleksandra Przegalińska, a philosopher and researcher of the development of new technologies.

RZESZÓW FORGOES ARTIFICIAL INTELLIGENCE IN COURTS

The decision was made by way of a referendum by the residents of Rzeszów and it should be considered as a victory for defenders of human rights, who have exposed the irregularities of the TemidAI adjudication system.

Rzeszów is famous for having the most digitalized and open government in Poland. Every citizen has an app through which he or she can participate in the budgeting process, consult local legislation, vote for candidates for boards of public companies and follow the actions of city politicians. All data aggregated in the city are open, which facilitates the introduction of new technological solutions. The TemidAI algorithm which, almost immediately after entering the data, was highly effective in resolving cases involving minor offences and selected cases of offenses committed on the Internet (due to the “digital” nature of the evidence) was the jewel in the crown of the open government in Rzeszów. 

TemidAI ruled without the involvement of a human judge on the basis of quantity-based norms, while taking into account the following three elements: (1) evidence entered in the system (videos, e-documents, recordings of interviews with witnesses); (2) analysis of historical judgments (the continuously updated database includes all judgments since 2030); and (3) analysis of the digital footprint of the parties and witnesses involved in the case. This significantly accelerated the resolution of the simplest cases and reduced the costs of litigation virtually to zero. Rzeszów also boasted about the effectiveness of the system measured by the lowest percentage of changed decisions in courts of second instance in Poland, where professional judges still decide.

The effectiveness of TemidAI was questioned by lawyers and representatives of non-governmental organisations, who pointed to a reversed causal process. The Helsinki Foundation for Human Rights (HFPC) has proven that human judges decided to uphold the judgments handed down by the algorithm in the first instance without a thorough analysis of the facts and evidence gathered. According to the HFPC, judges relied too much on the algorithm and were afraid to put their reputations on the line by questioning the “objective” resolution of advanced technology.

Meanwhile, instead of focusing on the individual aspects of the case, TemidAI made decisions on the basis of simple analogies, often by “stretching” the facts to match past decisions. Those who had had brushes with the law in the past had a disproportionately high probability. Moreover, TemidAI quickly generated straightforward correlations between, for instance, the level of wealth, place of birth and residence or grades given at school and the probability of committing an offence. The more decisions taken by the algorithm were entered into the judgment database, the more evident the irregularities described above became. This contributed to the creation of a system described by the HFPC as a “selffulfilling digital prophecy.” The foundation’s lawyers analyzed more than 8,000 cases and demonstrated that almost 6% of them had not had sufficient evidence to issue a judgement.

The decision made by the residents of Rzeszów is an important step in stopping the process of judicial automation – TemidAI was to be rolled out shortly nationwide. The opinion of the residents of Rzeszów will give a boost to opponents of such a change who, as the President of the HFPC said, “have tangible evidence that justice without the human factor can be simple, quick and cheap, but not just”. According to the Ministry of Justice, the abandonment of the new solution that the ministry has worked on over the last 5 years will slow down the technological revolution in the judiciary, but not stop it.

CONTROVERSIES OVER THE PRESIDENTIAL DEBATE IN THE US

“The Daily” podcast confirmed unofficial reports that the FBI intends to launch an investigation into a defective “truth algorithm” that did not work during the last Jindal – Obama debate.

It appears that the protests of Malia Obama’s team will be taken into account and the FBI will start an investigation to check whether any irregularities occurred when using the Real Time Fact-Checking system. According to the information revealed by “The Daily,” Chinese citizens are the owners of one of the six companies that create “truth algorithms.”

The reports have caused outrage among Republicans, who claimed the Democrats were sore losers. And the fact that the system detected only 14% of false statements (the average in previous debates exceeded 40%) shows that Bobby Jindal is “telling it as it is.” The CDO (Chief Digital Officer) of Jindal’s team emphasised that both teams had the same amount of time and the possibility to check the algorithms that determine whether the statements made by the candidates were in line with the facts and scientific knowledge.

The debate was extremely intense. Obama lost the topics related to China, climate, climate migration and e-war with Russia. She won the technological section and the section on minority rights. To a large extent, her failure was due to the high falsehood coefficient (the ratio between false and true statements), which was over three times higher than that of her rival. Many commentators were shocked by the result because Obama was famous for her excellent preparation and ability to phrase a message. Moreover, just after the winning debate, Jindal used his advantage and launched a social media campaign that perfectly matched his campaign slogan, “Telling it like it is”.

It is still unclear who will become the 64th President of the United States but according to experts, “Debategate” will influence the outcome of the election. Today, it seems that Jindal would have a great chance of winning, particularly after taking the “blue states” (supporting the Democrats) – Wisconsin, Georgia, Florida and North Carolina, or perhaps even Texas, which has consistently voted in favor of the Democrats since the election in 2032. If the immediate investigation by the FBI does not reveal any irregularities, the Republicans will gain an advantage and massacre Obama as a candidate who does not speak the truth and resorts to dirty tricks to hide it. And if the Democrats prove that irregularities have occurred, and China has actively supported Jindal, the Obama campaign will regain its vigour, which will most likely allow the president to secure her second term of office.

COMMUNITY DIALOGUES WILL HELP SELECT THE RIGHT NEIGHBORS

My Second Life employs DeliberateWe, a Canada-based company, to develop a transparent selection algorithm. Deliberations of individual communities will act as the starting point.

My Second Life (MSL), a leader in the market of luxury housing for senior residents (colivings), is introducing a new rating system (credit score) of future residents. This serves as an escape from accusations that the existing system, which is based on digital trace, largely assessed the profitability (for the company) of candidates for coresidents and did not appreciate their social virtues, which in turn may translate into group cohesion and a good atmosphere.

The previous system was supposed to take into account the habits and behavior of candidates for residents. A single algorithm was applied to all enclaves, irrespective of the nature and preferences of the communities living there. It consisted of the digital trace created on the basis of IoT (Internet of Things) readings in previous places of residence. It was also to take account of the “civic credit score” (developed on the basis of social interactions). However, this component, together with the assessment of the resident’s character, was underestimated and the algorithm primarily measured the “cost-effectiveness” of the future resident’s behavior and habits for the lessee. This led to conflicts between the older and the new members of the community who often followed a different set of values and consequently acted in a manner that was unacceptable to the communities.

Last year, MSL admitted that the algorithm was “suboptimal” and started searching for other automatic selection mechanisms. This week, it engaged DeliberateWe to hold dialogues with the residents of individual communities, thanks to which people interested will indirectly identify for themselves the guidelines which will form the basis for rating algorithms for individual enclaves.

Previously, the company had carried out a pilot program where the residents voted on the guidelines, but this solution did not work out. The residents of virtually all the enclaves tested voted primarily on those features and behaviors suggesting high sociability. For most communities, the recruitment procedures carried out on the basis of these guidelines resulted in a mismatch between the new residents and the communities’ praxeology, and in a decline in the neighborhood satisfaction ratio. This situation gave rise to the idea that each community should create its own algorithm itself and that this algorithm should be determined by experts holding so-called community dialogues, not by a simple vote.