We have all heard the fear-mongering predictions about AI, which has been accelerated by the adoption of automation during the pandemic. This extract from Phil Jones’ new book is drawing attention to the impact this rise in automation and AI has had on “microwork”, which despite being around for a decade is being used more frequently over the last few years and specifically with the advent of powerful crowdsourcing marketplaces like Amazon Turk.
The volume of microwork available via such sites has increased as AI has become more mainstream due to the need for human oversight, such as tagging objects in photos, to train algorithms and improve the AI. The work isn’t substantial enough for full time employees, so tech companies open it up on crowdsourcing sites to short-term freelance workers. And there are benefits to those that undertake such work, such as flexible working and additional income alongside other jobs, which harks back to the utopian model that companies like Uber and Deliveroo first set out around working when you want and as long as you want. However, whereas Uber and Deliveroo eventually succumbed to public pressure to introduce fair working practices, microwork has gone completely under the radar.
Concerns are now being raised about the creation of a dark job market and even a dark economy, in order to support the development and training of new and emerging technologies. This remote job market can and does exploit the most vulnerable, as these “ghost workers” are isolated and anonymous without a support structure to protect them.
The idea that humans are not just necessary but crucial in the development of automated processes and AI is an interesting but perhaps unsettling concept – many of us believe that AI is developed in labs by people in white coats surrounded by robots – but the reality is that like small children, machines need to be trained and taught by real people. And while that training might not be perceived as skilled worked, it is a vital part of the process for tech firms.
This is a rarely discussed and mainly an unseen by-product of automation and increased use of AI. We’ve all seen the headlines about robots stealing human jobs but this is shining a spotlight on the role of businesses and Big Tech in managing how AI is developed and all the people involved in the process. Having such a diverse group training AI and helping in the development of automation processes is a huge opportunity to identify and eliminate bias, yet despite all of these companies paying lip service to developing Ethical AI, this doesn’t extend to the working practices that they put in place to train their AI systems to deliver better results faster.
Microwork and access to it, is a massive opportunity to provide income to diverse workers which will only enhance AI solutions but it must be regulated and managed safely and responsibly so as not to take advantage of these workers or miss their vital contribution. Tech companies need to be proactive in setting an example and extending their ethical AI stance to include embedding safe working practices for those training their AI systems, rather than staying silent in the hope this quietly goes away.