AI-style exploitation, an alarming new trend
In a world where artificial intelligence is on everyone’s lips for their own reasons, I’ve often wondered whether the benefits it brings outweigh the risks.
Some legitimate but gimmicky concerns we often hear include whether superintelligent AI will wipe out humanity, take our jobs, or drive one species to extinction and create another. There is a possibility that it will become possible to create killer robots that do not wipe out species.
My concern is always, “How much does it cost?” And whether AI will impact the future of humanity, including the norms and values that have been the basis of our socio-economic model for centuries.
Above all, how much are tech companies spending on key elements of machine learning and AI? That’s the environmental footprint of megaprocessors in an era of trying to reduce global warming emissions. The cost of data collection and labeling, not to mention the cost of its consumption and cooling, is human labor intensive and costly. That could form a discussion for another article on its own.
The social impact of AI is not yet clear. However, because AI is based on data and large-scale language models, it requires extensive human input during the training phase. Big tech companies rely on service providers who employ independent workers, often from low-income countries, often working in less-than-perfect conditions and receiving little income. This has led some to warn about the creation of “digital sweatshops” and the exploitation of workers in developing countries.
Some warn of the creation of ‘digital sweatshops’ and exploitation of workers in developing countries
Mohamed Chebaro
Last year, Time magazine reported that Kenyan workers contracted to monitor ChatGPT text data for “toxicity” were paid less than $2 an hour and exposed to explicit and traumatic content. It was reported that he had not received any compensation for the incident.
Untold Magazine revealed last month that such workers were also being recruited from refugee camps. The Syrian refugees, some as young as 21, are recruited from camps in Bulgaria. After brief initial training in “digital skills” and training in English, refugees work part-time for data labeling companies, known as “microwork” or “clickwork,” which employers regard as trivial. I describe it as something simple. But the work they’re doing tells a different story.
These refugees spend their days labeling images of people based on race, gender, and age, performing something called “semantic segmentation” of satellite-sourced imagery. This is an important task for computer vision that involves careful separation and labeling. of every pixel in the image.
The report argues that this form of data work is particularly important in generating training data for AI, particularly computer vision systems embedded in devices such as cameras, drones, and even weapons.
The task at hand required workers to separate trees from bushes, cars from people, roads from buildings, and more. But experts argued that while such work is downplayed by employers as small-scale, low-skilled and requires no expertise, this is not actually the case. Some suspect that AI will be used for weapons and military applications because of the need for contextual knowledge, such as understanding what vegetation or vehicles look like in a particular area.
According to the World Bank, there are between 154 million and 435 million data workers around the world, the majority of whom live in the Global South. Their work is often done on a freelance basis, often commissioned by employment platforms or online companies, and they earn a few cents for each piece or task. They have no employment protection at all.
Their labor contributes to the development of algorithms that can cause discrimination and harm.
Mohamed Chebaro
In the case of Syrian refugees living in Bulgaria, their suspicions are not unfounded, as the number of autonomous drones and other similar technologies has increased dramatically in recent years. Tech companies and militaries are racing to integrate AI technology into their reconnaissance, target identification, and decision-making processes. Many are touting AI technology as a future war tool, although its use is clearly limited in the Ukraine-Russia war, and the broader Israeli war in Gaza.
Cheap labor is often the deciding factor in gigs and temporary freelance work, but the harm is compounded in the context of generative AI and machine learning models. Not only is this bad for low-wage workers, but their work can lead to discrimination and harm in ways that blur the lines of accountability and transparency regarding inaccurate or biased information. This is due to the fact that it contributes to the development of certain algorithms. Often the data comes from a third party or third party service provider.
One risk we are already experiencing is from AI-powered software tools used for surveillance and facial recognition, which have been shown to discriminate on the basis of race and gender.
Workers’ ethical concerns and moral dilemmas need to be addressed because some of their work can negatively impact their own communities. Many data label creators are secretive about their end users and their usage privileges.
If used in the right way, AI has the potential to be transformative and make a huge difference to our daily lives. But the lack of transparency of AI companies, such as the veil of secrecy surrounding the development of their programs in the name of protecting competitiveness, coupled with weak or non-existent government oversight, makes it difficult to believe that even if we believe However, it fosters a culture of impunity. There is an assumption that all companies operating in this field prioritize the best interests of humanity.
I’m still not convinced, especially when business ventures are primarily focused on profit and power. They are rarely established for the benefit of humanity as a whole. This is especially true for poorly paid workers.
- Mohamed Chebaro is a British-Lebanese journalist with over 25 years’ experience covering war, terrorism, defence, current affairs and foreign affairs. He is also a media consultant and trainer.
Disclaimer: The views expressed by writers in this section are their own and do not necessarily reflect the views of Arab News.
