The Hidden Human Cost Behind AI's Rapid Expansion and Ethical Failures

The Hidden Human Cost Behind AI's Rapid Expansion and Ethical Failures

The threat of artificial intelligence is not just real, it's here

The Hidden Human Cost Behind AI's Rapid Expansion and Ethical Failures

Generative AI and large language models are increasingly influencing daily life. These tools rely on millions of underpaid workers who label harmful content to train systems like ChatGPT. However, their development has raised concerns about the treatment of vulnerable communities and the fairness of AI-driven decisions. In November 2022, OpenAI launched ChatGPT, a chatbot powered by a large language model. To improve its systems, the company partnered with Sama, a US-based data-labelling firm. Sama recruited workers in Kenya, paying them between $1.32 and $2 an hour to review extreme online content. Many of these moderators later reported severe psychological trauma from the work. Sama has since faced significant backlash over its labour practices.

In the US, law enforcement has used facial recognition and predictive policing tools to target crime more efficiently. Yet these technologies have disproportionately impacted communities of colour. Some cities, including San Francisco and Boston, responded by banning facial recognition software due to its tendency to misidentify people unfairly.

Similar issues have emerged in Denmark and Australia. Both countries used machine-learning algorithms to assess communities’ eligibility for government benefits. In 2021, these programmes were paused after evidence showed they could produce discriminatory outcomes. The expansion of AI tools depends on addressing their impact on marginalised groups. Poor working conditions for data labellers and biased algorithms in policing and welfare systems remain unresolved challenges. Without solutions, these problems may persist as the technology spreads further.

Neueste Nachrichten