Progressive candidates and elected officials have a lot on their plates. Typically, they find themselves surrounded by corporate-controlled individuals with little to no interest in making genuine change for the public good. Developing progressive strategies is a collective process, one where we can all learn from existing studies and each other. This strategy review focuses on the implementation of artificial intelligence programs at the local, state, and federal levels, pulling from an extensive study by Stanford University .
Artificial intelligence is already integrating at the federal government level. The Stanford study details how AI serves five core purposes today: enforcing regulatory mandates centered on market efficiency; workplace safety; health care; environmental protection, and adjudicating government benefits and privileges, from disability benefits to intellectual property rights. An important application includes monitoring and analyzing risks to public health and safety, extracting useful information from the government’s massive data streams, from consumer complaints to weather patterns and communicating with the public about its rights and obligations as welfare beneficiaries, taxpayers, asylum seekers, and business owners.
Image Source: Stanford University
The report also notes that although the federal government is exploring these options, they are significantly behind the times—with only 12% scoring as high sophistication in their deployment. That’s not surprising. We have to compare them against the tools developed by knowledge economy companies such as Facebook, Google, and Amazon. Here we can identify an immediate concern of any progressive candidate. Without more funding and expertise, the government will fall far beyond the private sector in its capability and capacity to serve the citizenry. This lays the groundwork for the persistent and dishonest conservative talking point of inefficient government services that the private sector would perform better.
The first thing every progressive should consider is the ethical implications of implementing artificial intelligence into government systems. Specifically, which methods do we want to pursue, and what areas should we avoid? Currently, there seems to have been little to no public discussion about the ethical and human costs of widespread and unchecked AI implementation among mainstream progressives. For brevity, we’ll focus on exploring law enforcement, disseminating information and services, and determining why progressives should be learning and understanding the transformative potential of AI.
The most evident ethical conflict in our use of AI in law enforcement is especially relevant, given our present circumstances. Nearly 100 state and local jurisdictions have replaced traditional surveillance cameras with more sophisticated AI-powered gunshot detection technology. Others have employed AI-driven automatic license plate readers. Police departments in Los Angeles, Chicago, New Orleans, and Missouri have deployed AI-powered predictive policing strategies to identify gang-related crimes.
Machine learning algorithms enhance their predictive powers by processing existing connecting data points. There is ample evidence demonstrating racial bias in policing throughout the United States . Any predictive technology would only serve to accelerate and expand the racial bias because those are the data sets available. Justifying supporting this technology would require a total police audit, every precinct in all municipalities—unrealistic if manual but possible using AI. Without national police audits, the progressive must oppose predictive law enforcement technologies targeting individuals.
We also have to consider the bias occurring during the development of these algorithms. In a 2019 TIME article, computer scientist Joy Buolamwini shares that less than 2% of employees in technical roles at Facebook and Google are black. At eight large tech companies evaluated by Bloomberg, only around a fifth of the professional workforce at each are women. Her research also found a government dataset of faces collected for testing containing 75% men and 80% lighter-skinned individuals and less than 5% women of color. Unsurprising, if you consider that the majority of highly skilled computer developers are white males.
Highlighting these disparities isn’t a finger-pointing exercise in blame. It’s a recognition that our present circumstances have empowered one group over another. All of us bring a history of experience and perspective to our productive efforts, but in segmented professions like high-level development, many of these differences share similar themes. Therefore, until the field can diversify, we should all be wary of any predictive technologies using demographic information to determine criminality within the general population.
The use of AI should be given priority in the prevention of fraud in healthcare and finance even though its current form struggles to avoid narrow feedback loops and the dynamic change of fraudsters. These are crimes typically perpetuated by the wealthy. While they are often labeled “victimless,” sentencing is often a grim reminder of the existence of separate justice systems for rich and poor people. The problem facing many of our existing efforts is that they are not as automated as possible, creating efficiency gaps due to lack of resources. To develop active fraud and abuse detection, agencies need in-house experts to the same degree as any technology company would.
As is the case when implementing emerging technologies, the visionary leader’s objective is to see beyond what is. Progressive candidates and elected officials should be working towards reallocating AI funding from general law enforcement to fraud and abuse detection. The Trump administration has lifted the veil surrounding government spending. We can spend as much as we want whenever we want—there is never any lack of funds. It is always a matter of priority.
Here progressives may face significant challenges within the existing Congress. The suggestions we’re exploring relating to AI and law enforcement shift the predictive power of existing technologies away from citizen policing and towards the implementation of laws that impact a relatively small few, those with the means to commit massive fraud. Understanding that the majority of federally elected representatives are deeply beholden to corporations and wealthy individuals, we can assume that many representatives will attempt to stifle the enforcement of law onto these privileged groups.
Every progressive candidate and the elected official must develop the habit of raising a storm in the face of these injustices. The denial of investment to better track white collar crimes is a direct empowering of small groups to be above the law. When exploring AI in your policy platform, be sure to bring attention to those who oppose it—connecting their donors to potential conflicts of interest.
Progressives should also consider the challenge of how we regulate these efforts. Ideally, we want law enforcement to be as transparent as possible in all forms. We already understand the struggle associated with taking an opaque process and making it open, requiring our efforts to embed new policies with radical transparency. Unfortunately, this makes the struggle of predictive AI technologies more difficult. If we allow for things like the code and methodologies to be public, we run the risk of empowering organizations with the financial means to work towards developing methods to dodge fraud enforcement.
It is a paradox that requires careful consideration. We desire extreme accountability, but if we let the most well-resourced organizations work against the government, we will always be playing catch up. One alternative is to develop dramatic penalties for companies caught doing this. For example, we might legally require code audits for companies found in fraud to detect modifications made to go above the primary fraud parameters. Additionally, we could enforce penalties directly on CEOs and shareholders while incentivizing and protecting whistleblowers within the organizations. Progressives must reimagine frameworks for transparency and accountability beyond what history provides us, encouraging individuals to choose equal justice over elite favoritism.
The 2007 Census of Governments identified 39,044 general-purpose local governments, including counties, municipalities, and townships in the United States . One of the most significant opportunities for the enhancement of government services and impact is administrative artificial intelligence. Think about the many functions that the government manages and performs for us. We can significantly improve federal services like social security and corporate filings and registration, as well as local services like utilities and property taxes through the use of machine learning algorithms.
We are currently using artificial intelligence to process trademark and patent applications, disability claims, and postmarket surveillance of drugs by the FDA. In each of these applications, we can decrease the time associated with the process and successfully identify cases of interest. While challenges like flagging areas of interest incorrectly due to causal inferences based on unrepresentative data exist, they will most definitely improve over time.
Administrative AI contains the tremendous potential to transform the way people experience and understand the power of collective services administered by government agencies. The question every progressive candidate and elected official should be asking themselves is what areas of daily experience can be automated and improved through machine learning?
Public accountability is one area that can be dramatically improved. Imagine a community where the billing associated with property taxes, utility expenses, sanitation costs, and other miscellaneous projects are automated and adjusted in real-time. All bills are checked for errors and changed before being sent out, flagging potential issues (like a very high water bill) that are abnormal for review before delivery. The collection, processing, accounting, and communication concerning these aspects all handled automatically. Each new cycle allows the algorithms to better understand and predict potential issues within the community management. Eventually, a robust automated Q&A function will develop, pulling data and information from history to provide human quality (or better) feedback. Additionally, it eliminates repetitive jobs—quickly recuperating investment costs.
Administrative AI also allows for vastly improved crisis management. Imagine a scenario where a storm rips through a community, downing trees, and powerlines in the process. Towns with an AI reporting system could collect community feedback directly, continuously evaluating and categorizing incoming reports. Through preset rules and understanding, it can prioritize the areas of highest urgency, dispatch emergency crews with custom routes and time frames, communicate plans of actions with residents, and continually receive community feedback to reevaluate existing plans. What would take a mayor, council, and administrative staff days can process in real-time.
Forward-thinking communities can implement administrative AI for several innovative projects. Community resource networks such as aggregating energy purchases can use AI to monitor usage and automatically bill. Eventually, the AI will be able to offer predictive rate usage based on time and date, giving the municipality more substantial bargaining power when it comes to rate negotiation. Communities can invest in electric, autonomous vehicle fleets to serve their direct population. By offering free (driverless) rides to people within the city, we can help elderly and young residents navigate events and responsibilities more seamlessly than ever before.
Taxes are another project that could be handled entirely by artificial intelligence. Filing taxes could be free and straightforward, but H&R Block and Intuit are still lobbying against it .
Progressives must fight to end this corporate handout and allow Americans to enhance their tax experience actively. Aside from automatic filing, AI could be used to help identify and prosecute fraud. Presently the IRS is severely underfunded and targets lower economic class people for audits because it does not have the resources to fight and thoroughly investigate wealthy people . Administrative AI will provide pathways to ensure that the wealthy are held accountable to the fullest extent of the law while avoiding significant increases in labor costs currently associated with the task.
The Tenth Amendment to the U.S. Constitution grants all powers not given to the federal government back to the people through state and local government. The idea behind creating local governments is a concept allowing people to develop unique ways of living, experiments in being. Today we only see slight variations in a sea of homogeneity, which is precisely why every progressive should be pushing for investments in administrative artificial intelligence. After we address the basics, there is no limit to how customizable communities can become by utilizing the power of administrative AI.
Administrative AI provides a glimpse into a future of government services that perpetually improve themselves by continually learning, evaluating, and organizing our systems to be better and more efficient each year. With the support of dedicated technical teams, we could ensure that we address errors revealed during the process without losing sight of our end goal. The facilitation of government taxation and services should be automated to the highest degree possible. Doing so contributes to our capacity to improve upon and expand the programs we develop for ourselves.
There are several reasons why every progressive candidate should have at least a rudimentary understanding of the latent power artificial intelligence offers to government and citizens. More than anything, being progressive is a way of thinking about change. Nothing is going to be more impactful to the future of humanity than artificial intelligence.
Progressive candidates offer a new vision of the world, and all of us are approaching that from our unique perspectives. Many of us share an above-average technical and logical aptitude compared to the existing crop of representatives, which is why it is important to start conversations about automating reform now. Together we can build experimental programs at the local, state, and national levels to demonstrate the effectiveness and benefits of advanced AI applied to govern.
When we think about government in the long term of our lives, progressives should consider how we will structure it to undergo challenges and change regularly. The primary problem we have with government institutions today is that they resist change. Everything from property and contract to fraternal police unions exists to preserve the existing order. A strategy no longer valid in a world of exponential technological growth in nearly all directions.
There is no more significant opportunity for perpetual reform than to integrate administrative artificial intelligence into as many government service verticals as possible. By developing systems that can learn and change rapidly with circumstances, we overcome some of the most significant criticisms levied upon government—bureaucracy, and inefficiency. More importantly, we imbue challenge and change as the ordinary course of evolution in our social and legal arrangements.
We are at a unique inflection point in time that our children will not face. Major generational divides exist surrounding comfort and trust in emerging technologies between the millennials and Gen Z against the Baby Boomers. Where the boomers lack confidence and understanding, the younger generations willingly embrace advancement. Incorporating more AI solutions now will further empower future generations to make radical shifts for the collective good.
Whenever possible, progressives in state and local government should be fighting to dedicate resources to create in-house development staff. We cannot talk about administrative AI without realistically confronting the challenge, as mentioned earlier, of being beholden to the private sector. If elected leadership is unwilling or unable to establish publicly funded development teams, we can never be confident that organizational influences will not override the public good.
Technology is already ubiquitous within our lives and will only continue to be so, now is the time to expand public consciousness surrounding it. While most of the ideas here are evaluation and expansion of those mentioned in the sourced report, there are many other directions to consider. Education, energy, communication, the possibilities are endless. Progressive strategies could incorporate artificial intelligence programs now to lay the groundwork for a future that transcends our understanding of the possible.
 Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies by David Freeman Engstrom, Stanford UniversityDaniel E. Ho, Stanford UniversityCatherine M. Sharkey, New York UniversityMariano-Florentino Cuéllar, Stanford University and Supreme Court of California 2/2020 https://www-cdn.law.stanford.edu/wp-content/uploads/2020/02/ACUS-AI-Report.pdf
 What the Data Really Says About Police and Racial Bias Eighteen academic studies, legal rulings, and media investigations shed light on the issue roiling America.By Kia Makarechi 6/2016 https://www.vanityfair.com/news/2016/07/data-police-racial-bias
 Local US Governments by National League of Cities https://www.nlc.org/local-us-governments
 Filing Taxes Could Be Free and Simple. But H&R Block and Intuit Are Still Lobbying Against It. by Jessica Huseman ProPublica March 20, 2017 https://www.propublica.org/article/filing-taxes-could-be-free-simple-hr-block-intuit-lobbying-against-it
 IRS: Sorry, but It’s Just Easier and Cheaper to Audit the Poor by Paul Kiel ProPublica Oct. 2, 2019, https://www.propublica.org/article/irs-sorry-but-its-just-easier-and-cheaper-to-audit-the-poor