Cyber Democracy: Opportunities and Threats in Automated Lawmaking

While AI offers transformative solutions to longstanding democratic challenges, concerns about bias and cybersecurity linger. Can we balance innovation with integrity?

Good ol’ Democracy

As we’ve already discussed before the state of democracy isn’t the greatest worldwide, more and more authoritarian regimes are pulling off effective propaganda strategies positioning themselves as a way to provide well-being to their citizens (even though we know that to be a lie)But there is also a problem with the frustration that people have regarding their democratic regimes or elected representatives, weather is corruption, poor management overly bureaucratic systems democracy is not that well perceived by a lot of countries around the globe. 

And that is caused by many different factors, we live in an ever-accelerating world with radical technological breakthroughs happening every decade, now it's almost every year, and soon it will be almost every month depending on your industry of choice. But even in these exponential times, the user interface of democracy hasn’t changed much since before the industrial revolution. We go once every 4 or some years to a building we scratch a piece of paper an elected representative gets elected and their job (allegedly) is to represent the will of the people towards how to provide and develop policies and laws that will safeguard its people and promote welfare(oversimplifying a lot, of course) 

So what options and solutions or alternatives could we develop in the coming years to overcome these issues and give democracy a much-needed update where it could be a more efficient system of government that reduces people’s disenfranchisement while at the same time providing more and better solutions to its citizens and that can channel human capacity for problem-solving in an efficient manner to tackle the challenges humanity is facing in the XXI century. 

. 

The Threat of Misinformation Machines 

Another one of the most clear risks we discussed in a previous article, is the risk of misinformation spreading and accelerating thanks to the rise and pervasiveness of Large Language Models such as ChatGPT. These models are powerful tools that can exponentially increase the capacity of nefarious players to create convincingly and highly effective pieces of fake news or partially faked news that are designed for high virality.

We can see also the rise of generative AI platforms such as Midjourney (which I love so much) that can create accurate imagery that could be used as fake news or propaganda and that combined with LLMs could make any person with some descent coding skills into an automated fake news making machine with an incredible reach and devastating effects on the democratic and political discourse of a nation. So far we haven’t seen too much of this yet, but it is sadly an inevitability that these tools will be used in nefarious ways by agents of non-democratic regimes pretty soon ( assuming they aren’t already using them of course) 

So we need to be tackling this problem both from an education and digital literacy standpoint but also on a more structural robustness of our institutions to withstand the threat of things like these technological developments. Because a well-composed piece of fake news that is spread around enough becomes part of what a large percentage of people believe and the pervasiveness of fake news doesn’t discriminate through gender and education levels and income(though it has a deeper penetration in conservatives for some reason) it can attack everyone which is why it is so dangerous.

Digital Agents in our Democracy:

Let us get one thing clear, AI is already targeting you with political propaganda whether you’re aware of it or not, your feed in social media is tailored for your engagement, and the recommendation engines already know based on your likes, on your browser and search history what is your political tendency, so much so that not even the people who created these engines know(Hail the Shoggoth) 

So given the current state of affairs, the need to improve and enhance democracy is a powerful imperative, and even though civic education, and having a more engaged populace is a necessity. We might be able to incentivize more people to be a part of democracy if the system itself has a better and more responsive user interface where we as citizens can feel more part of the process and at the same time not feel overwhelmed by all the decisions that are required for a functional democracy. 

It is interesting seeing the proposal by Chilean Phycisits and author Cesar Hidalgo in his TED talk where he proposes the usage of software agents that are built from our political preferences and could handle in part some of the democratic decision-making where instead of having elected representatives we have these software agents as the members of this digital senate which has as many members as there are citizens and that could talk among themselves and in an automated fashion vote and handle the lawmaking process similar to how most congress and senates work worldwide.  

This is a very interesting proposition of automating lawmakers and elected representatives which are the bottleneck of the democratic process which sometimes can be a great thing because they are the buffer for heated emotional decisions that could arise in society from a stressful or highly chaotic period and put some cold water in the system when emotions run too hot. And don’t get me wrong I think this is actually a great thing because if we imagine a scenario of a direct democracy it is very naive because it assumes everyone who is capable of voting is educated enough about a certain topic or is also making that decision in a level headed way, which we know it is not true. 

AI's Role in Lawmaking:

While Hidalgo’s proposal covers a very interesting topic from a systemic point of view of integrating these automated representatives because they could act as lawmakers as well where they collect the needs and desires of the people in real-time through a myriad of methods extracting data from social media and analyzing the semantics and wording used and see the trends and needs of the people. Another method could be through the customization and tracking of user data and based on that proposes law and regulation. ( which has some issues regarding privacy, but then again is that even a thing that exists at this point?) Or simply have an online tool where every citizen can propose regulations and laws and if it has enough support it can be passed on to the automated senate to be voted or not. 

Now these systems could also be used on a smaller scale as an experiment with structures such as DAOs where an association of people, or a company is run by a computer program instead of people. Everyone in the group has a say in how things are done, and all the rules are recorded on a blockchain, so everything is transparent and no single person is in charge but everyone can see everything and check on the progress of the organization. And I think from DAOs there is a lot that could be learned and implemented in the democratic process mainly the transparency and also the capacity to create temporary organizations using this type of structure. 

Imagine the possibility of having a transparent and decentralized task force that could be easily structured created and funded using a DAO when a government has a crisis moment or when it needs to get a specific task that requires a more robust bureaucracy but not enough to justify an entire ministry or a new office within a structure like that. Providing more flexible and adaptable bureaucratic machinery of the state in service of the needs of the people.

Or having AI assistants to lawmakers who could instantly read proposed laws and offer a summary and other types of much-needed analyses shortening the time wasted at the moment of having interns and many experts around an elected official costing extra to the taxpayer. Not only accelerating and increasing the productivity of lawmakers but also avoiding conflicts with existing regulations or international treaties. 

Now there are of course concerns with automated lawmaking and DAO-like governments. On one hand, we can have automated systems enhancing and replicating inherent noxious biases of society towards a minority group and amplifying them instead of safeguarding the rights and freedoms of all of its citizens. 

And on the other, there are cybersecurity concerns because them ore these systems are sometimes centralized the risk becomes higher of the system being a target for hacking or other types of nefarious attacks. For example, if we integrated automated AI proofreaders for lawmakers if those systems are compromised maybe outside forces or political opposition could make these systems provide certain advice with a slight political bias towards their side for their opponents and it could sway the direction of some votes or how some new laws are designed through very sneaky subterfuge. Because we might think that an automated AI lawmaker is better than a corruptible human politician, but an automated politician instead of being susceptible to a bribe by a powerful force against a specific bill the entire system of automated lawmakers could be hacked and targeted, and then the entire system comes crashing down. ​​

Give me Cyber Democracy 

Now, personally, I WISH we could just automate and replace most congresspeople and senators in most democracies with AI systems that could in real-time gather information from the web, social media, and ourselves and based on that make informed decisions for the lawmaking process and help accelerate and reduce inefficiencies in the democratic process. In theory, it could be a great leap forward for our institutions and it would make us all feel more of a part of the democratic process no doubt. But, in reality, there are also key vulnerabilities in this type of system.

Coming soon to your nearest congress?

 

Because if a group is over-represented, for example on Twitter, then that group’s voice could be weighted heavier into the democratic process not because they represent a large % of the population but just because they are a loud minority. And these are the sort of things these systems are supposed to avoid. 

So hopefully a mix of better officials, automated “AI-Agents” senates, DAOs in government, and other structures are created and implemented in our democratic institutions in the coming decades so that way we can still enjoy the freedoms and benefits of democracy while the same time having elected officials (humans or not) and political institutions who are able to handle the coming challenges that we will be facing.