Data Privacy, Human Trafficking Risks on the Rise
The world can at times feel smaller and more connected than ever before. And though there are many tangible benefits to such intermingling of markets, people and products, potential risks nevertheless remain.
With greater connectivity comes the increased potential for breaches of private data. With bigger and more global workforces comes potential exposure to various forms of human trafficking. These are two risks insurers should be following as they have the potential to impact the industry as the year progresses.
The Increasing Threat to Personal Data Privacy
It’s no secret that data privacy has become a serious issue for many state legislatures – and a growing source of potential liability for companies. At least 50 bills addressing data privacy were introduced in 2023, compared to 34 bills in 2022, according to a Verisk analysis. Additionally, multiple states have enacted privacy-related legislation.
While “privacy” may be a fairly generic term, there are a variety of specific concerns that could fall within its orbit, such as biometric privacy, pixel tracking, and, increasingly, popular chatbots. In fact, more than 100 lawsuits have been filed since mid-2022 alleging privacy violations in connection with users’ interactions with website chatbots.
The popularity and potential privacy concerns associated with chatbots appear poised to increase in the coming years with the quick advancement of generative artificial intelligence (GenAI) tools. A little over a year since the public debut of one particularly popular chatbot, enthusiasm for GenAI among enterprise users appears to be continuing at a fever pitch. According to one survey published at the end of 2023, 67% of enterprises are using this technology and 18% already have applications in production.
Like many digital technologies, GenAI models appear vulnerable to cyberattacks. These attacks can occur whether a model is publicly accessible (i.e., via a chatbot on a website) or private (i.e., behind an organizational firewall). These exposures may come with a costly price tag for companies: By at least one account, the average cost of a data breach to U.S. organizations has been marching steadily higher – from $5.4 million in 2013 to $9.5 million in 2023.
If the growing adoption of GenAI applications is influencing the risk of data breaches, it may, downstream, potentially become a concern for insurers writing cyber or directors and officers (D&O) lines of business.
A Rise in Forced Labor Across Several Industries
Recent research and investigative reporting in the U.S. indicate that the record-breaking migrant crossings at the U.S.-Mexico border may be contributing to an increase in compelled labor across the country. A recent analysis by Verisk Maplecroft found a year-over-year increase of forced labor violations between the final quarters of 2022 and 2023, fueled in part by the recent influx of undocumented workers.
More than 27 million people are subjected to some form of forced labor, which, according to the U.S. Department of Homeland Security, can include work compelled through force, fraud or coercion. Forced labor generates an estimated $150 billion globally each year, making the practice the most profitable criminal activity in the world behind only drugs and arms trafficking.
A wide variety of industries may be potentially exposed, from mineral and resource extraction to cocoa and palm oil production to the training of large language models (LLMs) that power the increasingly influential AI chatbots offered by prominent tech companies.
A growing body of evidence suggests that some tech companies looking to capitalize on the fast-growing generative AI industry are harnessing the same model of digital supply chains to label and annotate data, a crucial component in training the LLMs that power AI chatbots and image generators. Demand for scalable AI training data has led to reports of exploited child labor in places such as Pakistan and Kenya – where workers were reportedly paid less than $2 an hour to train one popular AI chatbot to be less toxic.
The data-labeling industry is projected by some estimates to be worth anywhere between $17 billion and $70 billion by the beginning of the next decade. As this industry continues to grow, so too may the global demand for affordable – and exploitable – labor.
Data privacy and forced labor are complex issues exacerbated by the rapid advancement of GenAI. As the technology continues to gain traction, insurers need to closely monitor these associated risks.