● A team from the Civic AI Lab at Northeastern University (Boston) led by Saiph Savage has developed a suite of artificial intelligence tools to help gig workers for platforms like Uber and Deliveroo, among others, to identify and address specific socio-economic issues.
● Using new generative AI tools, they will be able to flag scams and payment defaults, and independently develop tailormade solutions like collective alert systems.
In the fast-growing gig economy, platforms like Fiverr, HopWork, Uber, Amazon Mechanical Turk, Superprof, and others offer piece rate pay for tasks like deliveries, taxi rides, and online classes etc. to “independent contractors”, who do not benefit from standard labour rights. Along with issues associated with their precarious status, these gig workers also face difficulties like constant surveillance by platforms, payment defaults, and in some cases, online scams. More often than not, they are too isolated to successfully deal with these obstacles. “If they had access to data from their platforms, they would be better informed about the problems they experience and better able to take collective action,” explains Saiph Savage, the director of the Civic AI Lab at Northeastern University (Boston). Using OpenAI’s application programming interface, Savage and her team have developed three artificial intelligence tools to make life easier for these unseen workers: GigSousveillance, GigSense and GigAction. “Most of the time, platform workers have no access to transparent job-related data.” GigSousveillance, for example, enables them to monitor those with authority over them (requesters and platforms) and collect data about their working conditions, which can be used to measure the extent of problems within a given professional context. For its part, GigSense allows them to interpret data on their working hours and evaluate their profitability, etc.
Generative AI helps these workers to structure and share their ideas, so that they can better respond to specific issues
Data visualization for a wider perspective on socio-professional problems
“GigSense provides a data-visualization tool that shows the scale of platform workers’ problems, which helps them to understand that they are not alone in experiencing these difficulties,” explains Kashif Imteyaz, a doctoral student at Northeastern University’s Khoury College of Computer Sciences. It enables them to collaboratively identify obstacles so that they are better able to deal with them. For example, individual freelancers who discover that they have responded to fake online ads would typically look to sites like Reddit for possible solutions, whereas with Gigsense, they can quickly join forces with other workers in a similar situation, and even go on to develop solutions like AI-powered web plug-ins to warn other platform workers. “In instances like these, generative AI can help workers to structure and share ideas, so that they can better respond to specific issues. Large language models enable them to take a step back and adopt a wider perspective,” point outs Saiph Savage.
Designing practical solutions with help from AI and human experts
Once a specific issue has been identified, gig workers who aim to develop a solution will need to design a mechanism for its implantation, which if it is to take the form of a plug-in will require a certain level of expertise. The GigAction AI model created by Savage and her team acts as a guide for workers navigating their way through the design process. At the same time, workers can also access advice from a voluntary community of human experts.
However, as Saiph Savage explains, further research will be required to perfect the new AI tools, which are just a first step towards promoting social responsibility: “Devising practical solutions is very complex, so we have obtained funding from the National Science Foundation to develop more resources that will enable workers to go further in understanding and resolving their problems.” Questions remain about the nature of AI models that are best adapted to empowering gig workers: should they be generic or highly specific systems focused on particular issues? However, given that an estimated 60% of the US workforce will be involved in the gig economy by 2027, it is clear that there is a growing need for initiatives of this kind.