Quest for transparency: Algorithms and gig workers' right to know

When an employee is disheartened with her boss, the first thing that she might do is speak to him; she may negotiate, may even quarrel. But what does one do when one’s boss is not a person but an algorithm? How can she question the decisions of her boss, if she cannot fathom how the boss decides? This is the situation of platform-based gig workers today.

Workers know that they are controlled by algorithms, but are unable to contest its decisions. Workers face material losses in their earnings but when approached, aggregators allege that the decision was not theirs but that of a machine’s. 

From reduced work allocation to being altogether blocked from accessing work on the platform, these systems exercise significant control over workers. The most attractive feature of platforms, when they first came to India over a decade ago, was that workers could ‘choose’ working hours. Flexibility and autonomy were key attractions for workers, making platforms a preferred source of income. 

However, algorithms have come to be designed based on exploitative incentive structures, which are developed by tracking the behaviour of workers.

Workers are allocated less work if they reject more than a permitted number of gigs, or they are incentivised to take more gigs to get bonus payments. The cut-throat rating requirements of aggregators have been a big bone of contention, with workers being arbitrarily ‘robo-fired’ from applications based on ratings, without being heard or given a reason, making flexibility a total myth.

Karnataka gig workers bill 

A novel attempt was made in Karnataka to bring transparency in algorithms of aggregators through the Karnataka Platform-Based Gig Workers (Social Security and Welfare) Bill, 2024.

This bill mandates that when sought by workers, aggregators must disclose information on the parameters for allocation and denial of work, categorisation
of workers, ratings system of the aggregator and the personal data of the worker being processed. 

Under the proposed law, aggregators are required to take measures to prevent discrimination by their algorithms on the grounds of religion, race, caste, gender or place of birth.

The bill also directs aggregators to appoint a human point of contact for workers to seek clarifications and mandates reasoned termination of contract with a fourteen-day notice period. 

Aggregators have deemed these compliances as intrusive on the grounds that algorithms are confidential and proprietary. They argue that sharing information on their algorithms would adversely affect competition and impede ease of doing business in the state.

There is no doubt that balancing the business interests of platforms against the information rights of workers is key. However, the answer is not in refraining from regulation altogether, but in determining and spelling out the level of algorithmic transparency that is expected from aggregators, which the bill currently does. 

The state government must certainly consider concerns on unwarranted disclosures of trade secrets, while formulating rules under this legislation.

The determination of the specific parameters of information to be disclosed by platforms must be a joint exercise, in consultation with both aggregators and workers. However, the argument that all information on algorithms of an aggregator is proprietary and confidential is far-fetched, overarching and vague.

This leaves no room for workers to understand the factors that influence their work. At the outset, the Karnataka bill is clear in defining the scope of the information that may be sought by workers, delineating that only algorithmic information directly affecting workers and working conditions are subject to transparency requirements. 

What more can the law do?

The Karnataka government’s attempt to bring transparency in the algorithmic management of work by platforms is not only laudable but also a first-of-its-kind initiative in India. 

However, the bill can benefit from adopting some of the progressive measures reflected in the Platform Work Directive adopted earlier this year by the European Union.

For instance, the directive provides that a person performing platform work cannot be fired or dismissed based on a decision taken by an algorithm, mandating human oversight on important decisions that directly affect workers.

The directive also prohibits platforms from processing personal data of workers to infer racial or ethnic origin, migration status, political opinions, religious or philosophical beliefs and trade union membership among other information. 

In line with the Platform Work Directive, the Karnataka law must also require a periodic assessment of the impact of algorithms on working conditions to be conducted by aggregators in association with workers’ representatives and labour experts.

The principal motive behind algorithmic transparency is to enable workers to collectively bargain for their rights by bridging the information asymmetry between workers and aggregators.

It is imperative that the proposed law must include provisions on the collective bargaining process to be adopted by workers going forward, in order to empower them to effectively negotiate pricing, work allocation, rating mechanisms and working conditions. 

Even as the strength of the gig workforce increases by the year, the debate on regulation has gone around in circles, sans any concrete measures to address issues raised by the workers.

With this bill, Karnataka has the potential to lead a nuanced regulation of platform work in the country. Workers must be empowered with information that rightfully belongs to them, and a law that mandates this is the need of the hour. 

(The writer is Senior Resident Fellow, Vidhi Centre for Legal Policy, Karnataka)

(This is the second of a two-part series on The Karnataka Platform-Based Gig Workers (Social Security and Welfare) Bill 2024.)