The parameters of this element are relatively clear. Career development can be experienced via intangibles like exposure and enrichment, as ladders to new jobs or better pay, and of course, as skills acquisition, either on the job or in formal training provided by platforms. What is more striking here is how much the experience varied across studies and contexts. Participants in these studies offered evidence for and against the experience of career development on each of these parameters.

The least tangible part of career development was exposure (or not) to new geographies, new cultures, and people outside networks. Freelancers performing transcription told Zollmann and Wanjala that they preferred work that allowed for personal growth (2020). Other microtaskers valued exposure to foreign clients (D’Cruz and Noronha 2016), or provided opportunities to carry out work they were unfamiliar with and experiences that they would not otherwise have been able to realize (Wood et al. 2019a). Others though, were not so sure. For example, there were drivers saying, “I don’t like this job. This is like a 24-hour service. There are health issues. There is no future here. I don’t like it. There is no personal growth.” Such drivers are aware of their inability to achieve growth in work or a change in occupation (Surie and Koduganti 2016, 28).

More tangible is career advancement or what some call “functional upgrading” (Graham, Hjorth, and Lehdonvirta 2017). Melia (2020) shared the story of one freelance writer in Kenya, who started as a night guard but became an online writer focused on bitcoin. He moved on to run an online business farming out tasks to other freelancers (a dynamic observed elsewhere; more on this on a crosscutting theme [Wood et al. 2019; Soriano and Panaligan 2019]). With more advancement, Melia explains how some online freelancers in Kenya do the work while in school, and then move on to white-collar professional work when the opportunity arises. Others who become subject matter experts stop working via platforms through word of mouth referrals (Melia 2020). 

Not everyone sees those paths, however. Freelancing offers no promotions, no certificates; nothing for the CV (Zollmann and Wanjala 2020). Work wasn’t even related to their area training (Muhindi 2019).  Gray and Suri (2019, xxv) describe microwork this way: “a revolving door of temporary tasks defines this job market. No obvious professional title. No ladder. No bonuses. No guarantees. Tasks are finite, built to disappear once a firm has reached its specific target and the people hired to hit it have moved on to other projects” (see also N. Gupta et al. 2014). The lack of portable ratings limits how well experience gleaned on one site can be translated to another (Melia 2020; Wood and Lehdonvirta 2019). 

Running a small shop with an online presence similarly may offer little upward mobility (Zollmann and Wanjala 2020) and in ride hailing, to exit a job is to exit the industry (Zollmann and Wanjala 2020).  Perhaps that’s why almost 60% of respondents in a survey of microworkers and freelancers expressed desire for non-crowd work, and 40% were actively looking (Berg 2016). It is also why “the relative novelty of online outsourcing makes it difficult to assess longer-term impacts such as those on career trajectories” (Malik et al. 2020, 15). 

Equally important is skill acquisition. Informally, freelancers can learn from family and friends (Wood et al. 2019b), trial and error (Margaryan 2016; Anwar and Graham 2019), or give opportunities to try new skills and interact directly with clients (Crosby and Cahaya 2017). Gray and Suri (2019, 110–111), note that microwork “can become a stepping stone or easy-to-access on-the-job training. On-demand work becomes a ‘sandbox’ where people can practice things like graphic design, typing, transcription, computer literacy, and language translation. These experiences are more difficult to acquire in a more traditional workplace, where there are greater expectations and pressure to perform.” They note that younger workers, and those with other revenue sources, may do microwork for skills acquisition instead of simply for revenue (Gray and Suri 2019).  

The other route to skills acquisition is via formal training provided by platforms. In previous work, we have described this as “platform-led upskilling” (Partnership for Finance in a Digital Africa 2019; Caribou Digital 2019). There are similar points raised in the various studies covered in this review. Raval and Pal (2019, 9) interview a manager of a local services place in India about the training the platform provides: “These pros eventually want to become entrepreneurs and we help them in getting there.” Uber and Cabify offer training in Cali (Reilly and Lozano-Paredes 2019). NGOs can partner with telcos to offer training to workers on platforms (Fairwork Project 2020). Even the World Bank funds training for platform workers in Pakistan (Malik et al. 2020). Through online training, workers gain interpersonal and business skills (Malik et al. 2020). Indeed, in many segments, it appears that guidance is a key determinant in whether growth is achieved (Genesis Analytics 2019).

Not all livelihood platforms provide training. Gray and Suri (2019, 80–81) note that “isolation and lack of guidance are spun as autonomy…. Workers do not know if they are technically or culturally competent for a task until they try to complete it. But tackling a job without knowing if they can succeed often presents a new risk…. If their reputation gets dinged, future job opportunities dry up.” In Kenya, Zollmann and Wanjala (2020) estimate that neither the transport sector nor the retail sector allows for formal learning. (For microwork and freelancing, see also N. Gupta et al. 2014; Uma and Marianne 2019.) 

Health & Safety: Are people healthy and safe when pursuing this livelihood?

This is an element where very little translates across different platform livelihoods. The threats, and occasionally the benefits, that accrue from a health perspective are largely a function of the vocation. Here are three illustrative contrasts about risks, where context and vocation are everything. 

It depends where the work happens. For example, in freelancing and microwork, working from home, often anonymously, can be perceived as relatively safe, perhaps particularly for women (McAdam, Crowley, and Harrison 2020). Other types of platform work, such as domestic help, are place-based, and may expose people, particularly women, to risks on early-morning commutes (Hunt et al. 2019), in uncomfortable and unprotected domestic situations (Aneja and Sridhar 2019), and to sexual harassment (Garcia et al. 2020; Kiarie, Singh, and Obiko 2020).

It depends on whether there’s cash around. One study argues that digital payments reduce the need for drivers in Indonesia to carry cash, and therefore reduce risk (Gachoka and Winiecki 2020), whereas drivers in South Africa became unsafe when cash option was introduced (Geitung 2017). Small retail merchants engaged in social commerce in Kenya (Partnership for Finance in a Digital Africa 2019) still operate in situations where the platform transaction is carried out in cash, a mode that feels risky.

It depends on how much risk the platform assumes. Contrasting platform driving with informal taxis, one study suggests that South African ride-hailing platforms provide accident and emergency medical insurance, making for a safer driving environment (Gachoka and Winiecki 2020). Another study of drivers in China notes that drivers don’t have these things (J. Y. Chen 2018). 

That said, there are two ways in which clear health risks have been identified. The first is around the physical safety of driving, which impacts both delivery and ride-hailing platform livelihoods. Motorcycle taxis in Kenya are not very safe (Zollmann and Wanjala 2020). Drivers in China report cutting corners, speeding, and ignoring traffic laws, in order to meet the deadlines imposed by the algorithm (J. Y. Chen, Sun, and Qiu 2020; B. Chen et al. 2020). Drivers in Indonesia report stress and exhaustion from riding long hours in traffic (Nastiti 2017). Drivers in South Africa don’t know where their ride will take them until they accept the fare, often exposing them to neighborhoods they perceive as risky (Geitung 2017). Indeed, in South Africa the risks don’t just come from carrying cash or speeding through red lights; it can come from the threat of physical violence from traditional taxi drivers, too. The considerable strife between platform drivers and taxi drivers in the country have been documented (Selabe 2017; Mare, Chiumbu, and Mpofu 2020; Geitung 2017).

The second clear risk is around the mental and emotional tolls of platform work. In microwork, this can include the harmful effects of watching pornography and violent content, if one is a content moderator (Gray and Suri 2019), and/or the grind of long hours and stress taking a physical toll on sleep (Anwar and Graham 2020; Wood et al. 2019a). That said, other studies suggest that health risks don’t appear prominently in the minds of platform workers (Muhindi 2019).

In general, it is difficult to disaggregate the risks associated with this type of work in general from the specific risks of doing this kind of work as facilitated and managed via a platform. What policymakers and the development community may have to look for is ways in which the particular affordances and constraints of platform work either reduce or accentuate the baseline levels of risk evident in each sector. One way to think about that is whether platforms do more or less than employers to mitigate risks, either through insurance, caps on hours worked, etc. or, as Gupta (2020) puts it, do they shift the burden of safety to workers? As an illustration, during the COVID-19 pandemic in 2020, some ride-hailing platforms provided personal protective equipment to their drivers, while others did not (Masiero 2020; Fairwork Project 2020). One could look at the relationships between the platform, the individuals, and the vendors plying livelihoods upon it through moments like this, and there’s no hard and fast assessment that would suggest that platform work is more or less safe than other types of work. It really depends on the context, the regulatory environment, and the choices platforms make. The opportunities for risks to be amplified or shifted to workers is discernible. It’s why groups like Fairwork are pressuring platforms to take more responsibility for the health and safety of their workers.