Introduction to the State of Remote Workers in the US

Since COVID-19, the world has become more adaptive and, I would argue, increasingly reliant on remote or hybrid work. When COVID-19 first emerged in March 2020, the number of Americans working remotely jumped to around 64%, according to a paper by Brannen, Ahmed, & Newton (2020). While the numbers have decreased since the height of the pandemic, remote work remains a significant part of the U.S. workforce. According to the U.S. Bureau of Labor Statistics, in 2023, about 27.8% of employed persons did some work at home on days they worked, and 15.2% of employed persons' work hours were at home (U.S. Bureau of Labor Statistics, 2024). With these high numbers of Americans working from home or in similar remote situations, many companies are exploiting workers and taking advantage of outdated labor laws. This issue is particularly prevalent in the field of AI training.

Legal Classification

A W-2 worker in the United States is protected by laws ensuring fair treatment and pay. Sadly, when labeled as a 1099 contractor, those laws do not apply, leaving workers vulnerable to mistreatment and misclassification. For example, companies often have terms and conditions requiring workers to agree to be contractors. By law, a contractor must be treated according to their title, meaning the hiring company does not control the hours, tools, location, or processes of the work. Contractors usually get paid by the project, not by the hour. Conversely, W-2 employees receive specific benefits like overtime, breaks, workers' compensation insurance, and unemployment benefits. Contractors must handle their insurance and other benefits independently. This arrangement saves companies money but leaves contractors, especially those new to the process, in precarious positions.

My Own Experience

From my personal experience doing task-related work and annotation for AI training, companies like Remotasks and Outlier have been a nightmare. They not only provide inadequate support for their workers but also blatantly exploit labor laws. They misclassify workers as contractors to save money while dictating every aspect of the work as if they were employees. In my experience, these companies control pay rates, payment schedules, tools, work environment, and even the required work hours. Such control negates the advantages of being a contractor while denying the benefits and protections of an employee. For example, in my last project with Remotasks, I was required to provide a resume and undergo an interview process—clear indicators of an employer-employee relationship.

Why This is Important

While tasks like surveys might be acceptable for extra cash, projects involving meaningful work that shapes revolutionary technology, such as AI, require time, energy, and often specific education. The quality of data used to train AI models like ChatGPT and Gemini must be high, free from bias, and accurate. However, the current exploitation of workers in AI training compromises the quality of this data. Workers are treated like sweatshop labor, easily replaceable without proper investigation into their qualifications.

The implications of these labor issues extend beyond just the workers themselves. Poor working conditions and unfair treatment can lead to lower quality training data, which in turn affects the accuracy and reliability of AI models. This could have far-reaching consequences as AI is increasingly integrated into critical systems across healthcare, finance, and other sectors.

Conclusion

To harness AI's positive potential, companies must treat their workers fairly, providing the necessary support and protection. This approach ensures that workers can focus on their tasks, knowing they are protected, which in turn guarantees the production of high-quality AI models. As it stands, contractors training these models face negligent and unfair treatment, which does not incentivize quality work.

Potential Solutions

To address these issues, several steps could be taken:

1. Updating labor laws: Legislators need to revisit and update labor laws to account for the realities of remote work and the gig economy, especially in emerging fields like AI.

2. Stricter enforcement: Regulatory bodies should increase scrutiny of companies misclassifying workers, imposing significant penalties for violations.

3. Worker education: Providing resources to help workers understand their rights and the differences between employee and contractor status.

4. Industry standards: The AI industry could develop and adhere to ethical standards for data annotation and AI training, including fair treatment of workers.

Opinion Section

In general, the United States has become a submissive group, manipulated by media stimuli, while corporations and governments enact ethically questionable acts to further their agendas. We must prevent AI from becoming another controlled media outlet, further oppressing humanity. While I am passionate about AI's potential, I would rather see it destroyed than misused to oppress society further.

Thanks for reading.