TECH4GOOD
A recent Washington Post article highlighting the presence of digital sweatshops in the country added a new dimension to the ongoing conversations about AI. This time, it is not about ethical issues, biases, regulation nor its use for spreading false information. The report focused on the thousands of young Filipino gig workers allegedly being exploited to support the booming AI business. Did the article paint an accurate picture of the actual working conditions or just focus on the exceptions?
There is now a growing demand globally for data services driven by the exponential growth of AI adoption. The AI models need huge volumes of accurate, precise, and legible training data to get them and machine learning algorithms to make unbiased and sound judgments. Even the best models, if fed insufficient or low-quality training data, will not deliver the expected results.
Data annotators and data labelers are involved in preparing labeled or annotated data for machine learning and AI model training. For example, in image recognition like those in the medical sector, they may mark objects in images, while in natural language processing, they may annotate text data. To train the algorithms for self-driving vehicles, the workers need to differentiate bicycles from trees. They also label images so the models would be able to determine whether it is a glass full of water or half-full of water.
According to Markets and Markets, a global marketing research company, the global data annotation and labeling market is projected to grow from US$0.8 billion last year to US$3.6 billion by 2027, at a CAGR of 33.2 percent. ScaleAI, featured in the Washington Post report, is an emerging data services startup reported to be a provider of AI data to leading global AI brands. Telework PH, based in San Rafael, Bulacan, is one of the leading Filipino-owned providers in the sector. A 2021 ILO study says that a vast majority of online freelance work is performed by workers in the Global South, almost half of which are in India and the Philippines alone.
The Washington Post report rang alarm bells among my colleagues in the local AI space. A number of them raised issues with the way it highlighted the bad side of AI saying it is adding to the further demonization of AI. This author was actually interviewed by the authors of the report where I said that “micro-tasking such as those done in digital sweatshops can not be the future that we want our young workforce to have.” Until such time that better job opportunities are made available to them, gig work can be a good option for them.
Online freelancing is thriving in the Philippines. In my own estimate, there are today more than two million Filipino online freelancers. Although some of them do it as a side hustle, for the majority, it is their principal lifeline that brings food to the table. Online freelancing is close to my heart. During my DICT days, my team and I initiated the Rural Impact Sourcing initiative which was meant to capacitate countryside talents to become digital workers. From Batanes to Tawi-tawi, we have seen how online freelance jobs have created economic activities where work opportunities are almost non-existent.
Bad press will not just be bad for the local online freelancing industry but for the Philippines in general considering its global-leading position as an IT-enabled services hub. Further research shows that many Filipino data annotators and labelers have been in the business for many years and are generally satisfied with their pay and working arrangements. Most of those interviewed are working for companies that are not mentioned in the report and say that the ability to produce quality output is something they had to really work on hard as pay is closely connected to the accuracy of their work. They had to initially work in internet cafes because most of them did not have laptops to use when they were just starting.
Global demand for data services is growing fast but it looks like there is a short supply of professional data annotators. This is a good opportunity for the Philippines to professionalize the data workforce, address the supply gap globally, and bring jobs to the countryside.
While the report may have brought to everyone’s attention the unacceptable working arrangements of some of the Filipino data workers, what also bubbled up is the glaring propensity of tech giants to close their eyes to how their providers are treating their workers. Regulating the industry may not be easy to do as most of the platforms are not even operating physically in the country. Remember that movie on Blood Diamonds? ([email protected])
(The author is an executive member of the National Innovation Council, lead convenor of the Alliance for Technology Innovators for the Nation (ATIN), vice president of the Analytics and AI Association of the Philippines, and vice president, UP System Information Technology Foundation.)