Ethical Considerations in Data Science: Ensuring Responsible Use of Data

In April 2024, the United States Congress passed a new legislative measure titled Protecting Americans from Foreign Adversary Controlled Applications Act (PAFACA), a bill that many news media outlets referred to as the “TikTok law.” At the heart of PAFACA is a provision that would ban the popular social network TikTok from operating in all 50 states and overseas territories, and the concerns fueling the ban reflect valid concerns over ethical data usage.

TikTok is a popular video and smartphone-focused social network with 150 million American users, most of them teenagers and even earlier members of Generation Alpha. The owner of the platform is Chinese technology giant ByteDance, a company with close ties to the Communist Party and high-profile government officials in Beijing. The PAFACA concerns are that the app presents a national security threat because of its massive personal and behavioral data collection practices.

The lawmakers who pushed PAFACA through the House of Representatives presented evidence of TikTok data turned over to the Chinese government during the 2019 pro-democracy protests in Hong Kong. Another point underscored by U.S. lawmakers is that ByteDance does not seem to follow any responsible data practices that would prevent the Communist Party from demanding access to TikTok data arbitrarily. There is no evidence that this has happened to American users, but the PAFACA legislation was passed with bipartisan support.

There’s a vital lesson for brands and businesses to learn about PAFACA and what could happen to TikTok if ByteDance is forced to sell its American operations, which were valued at $100 billion in May 2024. Any company handling personal information or collecting behavioral data from customers must observe and implement measures conducive to data privacy ethics. TikTok is not the only major technology platform that has made news headlines in this regard; the Cambridge Analytica and Facebook scandal of 2018 was strongly linked to the controversial 2008 U.S. presidential elections, and ChatGPT is facing legal challenges concerning the General Data Protection Regulation (GDPR) of the European Union.

Understanding the Ethical Dimension of Data Science

The previous examples of questionable use of personal and behavioral information illustrate why data science ethics must not be ignored. As a scientific discipline and a professional field, data science is subject to certain codes of conduct and ethical practices. The ethical dimension of data science focuses on the responsible use of information collected, stored, and processed throughout the entire data lifecycle. Applying ethical principles ensures that data science practices are not only effective but also fair, transparent, and accountable.

The Need For Data Privacy and Protection Measures

Individuals have the right to control their data. The extent of this right varies by legal jurisdiction, and the United Nations has counted 137 of its 194 member nations with laws that guarantee the right to data privacy. Most of these laws have been passed since the aforementioned Cambridge Analytica scandal was exposed by international news media outlets. By this right, data scientists should obtain informed consent before collecting and using data; moreover, they should ensure that data is anonymized or pseudonymized whenever possible, which should be most of the time.

The obvious need for data privacy measures boils down to compliance. It is the main reason why we are seeing more business owners and brand managers request data scientist consulting services these days. Beyond compliance, however, data protection goes beyond simply respecting individual privacy; it is also essential for building trust and fostering strong business relationships.

Responsible Data Governance and Ethical Practices

Many of the ethical practices linked to data privacy can be achieved through established data security measures. When going deeper into data analysis, however, a layer of fairness should always be applied. Algorithms and data models should be unbiased; they must avoid discrimination based on factors like race, gender, or socioeconomic background. Data scientists should actively identify and mitigate potential biases in data and algorithms, particularly now that we are entering a new tech era of artificial intelligence (AI) constructs such as ChatGPT.

Transparency and accountability should be part of data governance. The methods and processes used in data analysis should be clear and understandable. Users should be informed about how their data is being used and for what purposes. Professionals who use data science tools should be accountable for the outcomes of their work; this level of accountability should include addressing potential harms caused by algorithms and data analysis. Responsible AI practices and AI transparency standards are just beginning to be considered, but they are already shaping new trends in data governance.

The Growing Importance of Ethical Guidelines in Modern Data Management Practices

Data management is a business process with a substantial potential for for both positive and negative impacts. When used responsibly, data can drive innovation, improve decision-making, and benefit society in numerous ways; we know this from the early days of the Digital Age in the mid-20th century. We also know what can happen when data is used irresponsibly; that is how we ended up with the Cambridge Analytica scandal.

As long we continue to guide our decisions by how we collect and analyze data, ethical guidelines should be observed and heeded by everyone involved in the data management continuum, which runs from the collection point to the analysis and decision-making stages. There is a lot to be done in this regard; for example, algorithmic bias can perpetuate existing societal inequalities, so data scientists should identify and mitigate bias in data collection, analysis, and model development.

Any decisions driven by data analysis must be fair and non-discriminatory. This is of particular importance when using AI constructs. There must also be a level of trust developed with the individuals from whom data is collected, and this is where transparency and accountability come into play.

Ann Mazotta, California Business Journal