Team AI Divisional Framework

In late March 2021, U.S. Representative Susan DelBene (D-WA 01) introduced data privacy legislation to the 116th Congress to protect consumer data from business malpractice. With data privacy legislation only existing in the U.S. at the state level (as of writing), this proposal introduces a novel idea of a federal data privacy policy that can provide standardized protection for Americans across all states.

Data privacy entails the protection of personal information from collection and analysis by businesses’ AI tools. At one point in time, data privacy online was viewed as a luxury, but with the increasing grasp of technology in our lives today plus numerous data privacy scandals in recent years, data privacy is now a necessity. People value their personal data and privacy from the reach of businesses who try to collect them for profit, and a new mindset views data privacy as a personal right as opposed to it being a privilege.

The AI team aims to address the lack of legislation regulating data privacy and collection by proposing the forecasting question “Will the U.S. government implement policy to regulate AI via consumer data protection from businesses in the current administration?” To derive our forecast to answer this question, we developed five reference classes to categorize and focus our case study research. This article is an in-depth analysis of each of our five reference classes and how important each one is (percentage-wise) in answering our forecasting question.

Public Concern Over Data Usage (5%)

Privacy is a fundamental human right recognized in the UN Declaration of Human Rights, the International Covenant on Civil and Political Rights and in many other international and regional treaties. Privacy underpins human dignity and other key values such as freedom of association and freedom of speech. It has become one of the most important human rights issues of the modern age. Within recent years and growing threats to data privacy, the framework and opinions towards data laws and protection have also adapted this view, viewing privacy as a right rather than a privilege, especially when it comes to personal digital data.

This is also not the opinions of a minority group or a small fraction of the U.S or even the world. 79% of U.S citizens do not trust the government with their personal information and 83% say the same for businesses. This distrust within the people against government and businesses is quite apparent and established. Even with countries who are the highest rated in data privacy and protection due to their continuously updated laws and regulations, the citizens of this country label data privacy on average one of the top five most valued rights within the country. (Survey taken from Finland and Switzerland). So in either case, people value their privacy especially when it comes to digital privacy and protection.

Businesses’ Influence in Pushing Policy (15%)

To understand businesses’ influence in pushing policy for or against AI regulation, we must first understand what role AI plays within businesses and why this role is so crucial to their function. A business’s main goal is to generate a profit for their company-- AI furthers this purpose by targeting customers, increasing efficiency through predictive analytics, and protecting against fraud schemes. Essentially, AI plays a crucial role in increasing a company’s productivity and tailoring products or services towards a target audience.

Now, how much influence does a corporation have with legislators? According to The Conversation, “Wealthy individuals and organisations have a disproportionate influence over elected representatives”. In 2018, Google spent $21 million on lobbying Washington. Power is concentrated in the hands of wealthy individuals who can move money around the globe untraced. This money is used to purchase the services of lobbyists or donate to political campaigns illegally, a privilege not granted to smaller businesses.

Let’s put the two together-- how will big corporations influence legislators to regulate AI and what are their motives in doing so? Sundar Pichai, CEO of Google, writes in his article that, “Companies such as ours cannot just build promising new technology and let market forces decide how it will be used. It is equally incumbent on us to make sure that technology is harnessed for good and available to everyone. Now there is no question in my mind that artificial intelligence needs to be regulated. It is too important not to. The only question is how to approach it.” Going back to the statistic of Google spending $21 million on lobbying, if Pichai wanted to regulate AI through backdoor means, he definitely has the power to. In economics, rules and regulations boost market leaders; Google is a market leader in the technology industry. Governments’ legislation will create further barriers to entry for competition in the market, which will ultimately benefit large corporations.

This is just one case of corporations influencing legislation through seemingly selfless advocacy or backdoor lobbying. Facebook, Cambridge Analytica, and Nestle are all examples of other companies accused of lobbying. The amount of illegal activities that go on behind closed doors would take years to fully research-- however, this case study proves that Google, just one of the tech giants, sets an example for other companies who act similarly. The passing of legislation regulating AI, natural language processing, and predictive analytics is highly impacted by corporate lobbying.

Malpractices of Companies (10%)

Oftentimes, delegates in Congress are introduced to issues regarding technology via investigations conducted by some other entity. This entity is often the Federal Trade Commission (FTC): an independent US government agency whose role is to enforce antitrust law and provide consumer protection. When entities like the FTC often find themselves exhibiting a trend in rulings/findings regarding multiple investigations and see a need for standardization across all businesses, their investigations can lead to congressional legislation.

For example, in the 1990s, e-commerce was on a sharp rise, but concerns regarding data collection methods and user privacy were growing, especially in relation to children, as few web pages had published privacy policies. After a petition by the Center of Media Education, the FTC investigated websites that catered to children, most notably being KidsCom.com. FTC found that KidsCom.com had engaged in ‘unfair/deceptive’ practices and published a letter indicating that the website’s data collection processes were susceptible to legal action. These events resulted in a determination of the need for a means to mitigate childrens’ online privacy risks and involve parental consent. Shortly thereafter, the Children's Online Privacy Protection Act of 1998 (COPPA), which applies to digital data collection regarding children under 13 by entities under U.S. jurisdiction, was introduced and passed into law.

The FTC has recently exhibited a similar attitude, as in late 2020, they “issued orders to nine social media and video streaming companies, requiring them to provide data on how they collect, use, and present personal information.” Thus, investigations into malpractice within these tech companies may have a critical effect on the likelihood of new legislation regarding consumer data.

Political Party Cooperation (60%)

When discussing the ability for legislation to be effectively passed, it is imperative to consider the impact of political party cooperation, and, on the other side of the coin, political party division. If political parties are polarized and divisive, it may seem as if legislation may never get passed; however, the truth is quite surprising. Researchers from Johns Hopkins University found that the passage of legislation had little to do with the polarization of the legislative or executive branch. Rather, it depended on the coalitions within Congress. Because elected officials are more likely to take a side based on issues rather than perfect party alignment, if a particular topic is of interest to certain legislators or to their constituents, it will be supported. Take for example the case of Texan legislators. Though the democratic representatives would not want to support the oil industry, they understand that Texas is one of the hotspots for oil production in the United States and will side with Republican legislators when it comes to preserving the oil industry (lest they be accused of contributing to growing unemployment in Texas’s oil industry).

In addition to partisanship, legislative inaccuracy plays a role in determining the passage of legislation. When polarization is severe and coalitions are unable to get the ball rolling on legislation. The legislation that ends up being passed tends to be very inaccurate in its wording and frequently contains many errors that must be addressed at the judicial level. When this occurs, the power to influence legislation leaves the hands of representatives and ends up in the hands of the appointed judges in the Supreme Court. Legislators obviously dislike this idea, and they are more likely to not pass legislation at all than risk the intricate wording of the legislation and be at the mercy of the justice system.

Hence, to fully consider the likelihood of AI-regulating legislation being passed by the US Federal Government, it is important to understand the issue of partisanship and legislative inaccuracy. Both those reasons have a significant impact on the likelihood of legislation being passed, and whether other parties such as people or businesses will affect decision making.

Incorporating Technology into Business (10%)

Over the course of the last two centuries, innovation has brought about new technologies like humans have never before seen. As a result, businesses have generally sought to incorporate new technologies and practices into their products and services to reduce costs and increase efficiency. From the Industrial Revolution to today’s technological revolution, humans have constantly found ways to expedite their work, but do these innovations have unintended consequences? And if so, do the benefits of incorporating technology outweigh the costs?

The benefits of technology in business are more evident than its drawbacks, as specifically AI technology has allowed companies to collect data to produce personalized advertisements, automate processes (like social media feeds, etc.), and reduce the need of human resources by replacing them with cheaper and more efficient computer algorithms.

However, these supposed benefits also bring the risk of misuse and exposure of personal information. In recent times, the most notable critique against the use of AI in business is how data collection has been abused to sell personal information without consent. A notable example would be the Cambridge Analytica scandal during the 2016 US Presidential Election where developers abused the Facebook application programming interface (API) to gain personal information about Facebook users which they sold without consent to target political advertisements. The issue of data collection exists outside the realm of social media as well, as inaccurate data collection to train AI algorithms also results in societal failures. For example, improper data training for AI facial recognition models has led to a bias in facial recognition technologies against people with darker skin tones. Additionally, Amazon has recently faced scrutiny for a resume filtering AI model that discriminated against women because it was fed with information about Amazon’s previous hires, which were largely male due to the male dominance in the tech industry.

Whether the use of AI has brought about more good or harm is a gray area that is up for debate, but one aspect that humans should be wary of is placing their full faith in AI; with improper training, AI can also inherit these same biases.

Unfortunately, at least in the United States, legislation has not kept up with the issues that AI has unveiled in the personal data sector. There is no federal legislation as of yet that deals with data privacy and only a handful of states have passed a form of data protection for their residents (notably California, Nevada, and Maine). Even though there is a trend of data privacy legislation passing at the state level, no federal standard exists for US businesses to comply with, leading to fragmented legislation and policies across state lines leading to confusion for both businesses and consumers.