What we can learn from Michael Jordan and the Chicago Bulls about data stewardship

Michael Jordan will shed light on personal data stores; we’ll use the full Bulls team to illustrate data cooperatives, and coach Jackson will help us describe data trusts.

The following piece was first published on the Data Empowerment blog and was originally published on May 17, 2021.  

Winning six championships between 1991 and 1998, the 1990s Chicago Bulls were one of the most successful teams in US sports history. Led by basketball legend Michael Jordan and coached by Phil Jackson, no team has yet topped the Bulls’ 1995–96 NBA season record of 87 wins vs 13 losses.

What does this have to do with data empowerment? Here we’ll use the 1990’s Bulls to explore three emerging mechanisms for data governance, or, more accurately, data stewardship: 1) personal data stores 2) data cooperatives and 3) data trusts. Each of these mechanisms offer an answer to one of the biggest questions of our time: how do we make data work for us?

Michael Jordan will shed light on personal data stores; we’ll use the full Bulls team to illustrate data cooperatives, and coach Jackson will help us describe data trusts.

Personal data stores

Personal data stores enable users to decide who their data is shared with and for what purposes through granular permissions, rather than data being collected, stored and managed by a company such as Google or Facebook. These provide both a storage mechanism for users’ data and the ability for users to control how third parties access this data. Examples of such technologies include Solid, Digi.me and Meeco, with many more personal data stores and services being launched all the time.

Personal data stores are for the Michael Jordans among us. They provide significant individual control and can be extremely useful for those with the capital, skills and time required to make informed decisions about how their data is handled, what it’s worth and how to make the best use of it. However, most people are not equipped, willing or interested to take that level of decision-making about questions over which few people (if any) are able to fully predict the outcomes. What’s more, we should remember that the individual decisions we make about ‘our’ data can impact others, much like the decisions that Michael Jordan took during close games affected his team and the entire franchise, for better or worse. Having full control over your data using a personal data store is both a privilege and a burden.

Data cooperatives

While personal data stores might be suited for the well-informed and experienced among us, cooperatives lend themselves to those who want to benefit from sharing and pooling their data with others and taking collective decisions about how it is used. Examples include Driver’s Seat and MIDATA and Carbon Co-op.

In a data cooperative members manage their data together, sharing it with each other and deciding if and under what conditions third parties can access it. Unlike personal data stores, data cooperatives shift power to individuals by forming groups that are better equipped to advocate for the rights of their individual users. The logic is that the interests of individual members can be better served in a collective than in conventional (hyper-)individualistic models.

As a team, the Bulls were successful not only because they had Michael Jordan in their ranks, but because they formed a strong collective that knew how to work together and make difficult choices as a team. The team would often shift the burden of responsibility from one single player to leverage the combined strength of the team making choices together. For instance, the team’s Triangle Offence required discipline and sacrifice from each player rather than the expression of their individual talent alone. It was brutally effective.

While a data cooperative requires a certain level of decision-making over their data, the collective pooling of data increases the benefits of each individual as the power of a group grows with each cooperative member.

Data trusts

The currently much-hyped data trust concept takes the idea of data stewardship to a whole new level. In a data trust, one or several trustees are responsible for stewarding someone else’s data or data rights in the sole interests of that person. While personal data stores are based on a contractual relationship, data trusts rely on a legal relationship between the data subject and the data steward where the trustee has a fiduciary duty to act according to predefined terms and conditions. The data trust concept is fairly new and so examples remain limited.

The trustee of a data trust is similar to a coach of a basketball team — albeit with even more control. Though Phil Jackson would often count on his team to make the right choices, when his Bulls weren’t able to come up with a solution, he would call a timeout and take decisions on the team’s behalf and in the players’ best interests. Data trusts work in this way all the time, with the burden of complex data decisions transferred from individuals to specialized intermediaries.

Making the right decisions over someone else’s data should, however, not only depend on the integrity, honesty and capability of the trustee but needs to be guaranteed by an underlying regulatory framework. Given the inherent risk in relying on someone else to make decisions in one’s best interest, the data trust members need to be certain, and have a legal guarantee, that their trustee will act with “prudence, impartiality and undivided loyalty on their behalf”.

Differences and Similarities

Which of these models is best suited to give people (more) control over their data? Let’s look at the similarities first:

All three seek to address power imbalances between ‘data subjects’ and data controllers, and limit the increasingly concentrated power that our data gives the companies that own the infrastructure, technology, apps and services we use. All three give individuals (and groups) authority over the data they produce. All serve their users (members), which is very different from the current model where companies like Google and Facebook have a duty to protect the interests of their shareholders above those whose data they collect. Finally, all three try to do so by decoupling data collection, management and control from data use.

The differences:

Personal data stores focus on the individual, giving individual users effective and consistent control over their data and decision-making power over its access and use. Thus, individuals are able to extract value from their data, including monetary gains.

In a member-controlled data cooperative, individuals give up a degree of decision-making power so that decisions about the collective’s data are made democratically through direct or delegated voting. Data cooperatives are often formed between like-minded people with a shared economic, social or cultural interest that can be achieved more easily by voluntarily pooling their data.

While data cooperatives can be used to facilitate the pursuit of societal goods (e.g. cooperative members sharing their anonymized data for health research), data trusts, where trustees act as intermediaries and are bound by a fiduciary obligation to its members, might be better suited for the creation of societal value. Knowing that a regulatory framework underpins a data trust which requires trustees to act in the interests of the beneficiaries and not in their own self-interest, people might be more inclined to hand over some of their data (rights) to a trust that can then share it with third parties for societal or environmental purposes, thereby fulfilling the idea of data as a common good that benefits society as a whole.

The way forward: the many, not the few

We’re not advocating for the supremacy of any of these three models — they can all be effective in different circumstances. Just as we need teams, star players and effective coaches, we need various tools to control, steward and leverage data in a way that best serves us as individuals and society at-large.

In future posts we’ll discuss which of the models offer the most promise under certain circumstances and contexts.

— — — — — — — — — — — — — — — — — — — — — —

Further reading

Check out these resources if you want to dive deeper into data stewardship:

Ada Lovelace Institute & UK AI Council: Exploring legal mechanisms for data stewardship

Aapti Institute: Data stewardship — a taxonomy

Mozilla Foundation: Data for empowerment

The following piece was first published on the Data Empowerment blog and was originally published on May 17, 2021.  

 

Winning six championships between 1991 and 1998, the 1990s Chicago Bulls were one of the most successful teams in US sports history. Led by basketball legend Michael Jordan and coached by Phil Jackson, no team has yet topped the Bulls’ 1995–96 NBA season record of 87 wins vs 13 losses.

What does this have to do with data empowerment? Here we’ll use the 1990’s Bulls to explore three emerging mechanisms for data governance, or, more accurately, data stewardship: 1) personal data stores 2) data cooperatives and 3) data trusts. Each of these mechanisms offer an answer to one of the biggest questions of our time: how do we make data work for us?

Michael Jordan will shed light on personal data stores; we’ll use the full Bulls team to illustrate data cooperatives, and coach Jackson will help us describe data trusts.

Personal data stores

Personal data stores enable users to decide who their data is shared with and for what purposes through granular permissions, rather than data being collected, stored and managed by a company such as Google or Facebook. These provide both a storage mechanism for users’ data and the ability for users to control how third parties access this data. Examples of such technologies include Solid, Digi.me and Meeco, with many more personal data stores and services being launched all the time.

Personal data stores are for the Michael Jordans among us. They provide significant individual control and can be extremely useful for those with the capital, skills and time required to make informed decisions about how their data is handled, what it’s worth and how to make the best use of it. However, most people are not equipped, willing or interested to take that level of decision-making about questions over which few people (if any) are able to fully predict the outcomes. What’s more, we should remember that the individual decisions we make about ‘our’ data can impact others, much like the decisions that Michael Jordan took during close games affected his team and the entire franchise, for better or worse. Having full control over your data using a personal data store is both a privilege and a burden.

Data cooperatives

While personal data stores might be suited for the well-informed and experienced among us, cooperatives lend themselves to those who want to benefit from sharing and pooling their data with others and taking collective decisions about how it is used. Examples include Driver’s Seat and MIDATA and Carbon Co-op.

In a data cooperative members manage their data together, sharing it with each other and deciding if and under what conditions third parties can access it. Unlike personal data stores, data cooperatives shift power to individuals by forming groups that are better equipped to advocate for the rights of their individual users. The logic is that the interests of individual members can be better served in a collective than in conventional (hyper-)individualistic models.

As a team, the Bulls were successful not only because they had Michael Jordan in their ranks, but because they formed a strong collective that knew how to work together and make difficult choices as a team. The team would often shift the burden of responsibility from one single player to leverage the combined strength of the team making choices together. For instance, the team’s Triangle Offence required discipline and sacrifice from each player rather than the expression of their individual talent alone. It was brutally effective.

While a data cooperative requires a certain level of decision-making over their data, the collective pooling of data increases the benefits of each individual as the power of a group grows with each cooperative member.

Data trusts

The currently much-hyped data trust concept takes the idea of data stewardship to a whole new level. In a data trust, one or several trustees are responsible for stewarding someone else’s data or data rights in the sole interests of that person. While personal data stores are based on a contractual relationship, data trusts rely on a legal relationship between the data subject and the data steward where the trustee has a fiduciary duty to act according to predefined terms and conditions. The data trust concept is fairly new and so examples remain limited.

The trustee of a data trust is similar to a coach of a basketball team — albeit with even more control. Though Phil Jackson would often count on his team to make the right choices, when his Bulls weren’t able to come up with a solution, he would call a timeout and take decisions on the team’s behalf and in the players’ best interests. Data trusts work in this way all the time, with the burden of complex data decisions transferred from individuals to specialized intermediaries.

Making the right decisions over someone else’s data should, however, not only depend on the integrity, honesty and capability of the trustee but needs to be guaranteed by an underlying regulatory framework. Given the inherent risk in relying on someone else to make decisions in one’s best interest, the data trust members need to be certain, and have a legal guarantee, that their trustee will act with “prudence, impartiality and undivided loyalty on their behalf”.

Differences and Similarities

Which of these models is best suited to give people (more) control over their data? Let’s look at the similarities first:

All three seek to address power imbalances between ‘data subjects’ and data controllers, and limit the increasingly concentrated power that our data gives the companies that own the infrastructure, technology, apps and services we use. All three give individuals (and groups) authority over the data they produce. All serve their users (members), which is very different from the current model where companies like Google and Facebook have a duty to protect the interests of their shareholders above those whose data they collect. Finally, all three try to do so by decoupling data collection, management and control from data use.

The differences:

Personal data stores focus on the individual, giving individual users effective and consistent control over their data and decision-making power over its access and use. Thus, individuals are able to extract value from their data, including monetary gains.

In a member-controlled data cooperative, individuals give up a degree of decision-making power so that decisions about the collective’s data are made democratically through direct or delegated voting. Data cooperatives are often formed between like-minded people with a shared economic, social or cultural interest that can be achieved more easily by voluntarily pooling their data.

While data cooperatives can be used to facilitate the pursuit of societal goods (e.g. cooperative members sharing their anonymized data for health research), data trusts, where trustees act as intermediaries and are bound by a fiduciary obligation to its members, might be better suited for the creation of societal value. Knowing that a regulatory framework underpins a data trust which requires trustees to act in the interests of the beneficiaries and not in their own self-interest, people might be more inclined to hand over some of their data (rights) to a trust that can then share it with third parties for societal or environmental purposes, thereby fulfilling the idea of data as a common good that benefits society as a whole.

The way forward: the many, not the few

We’re not advocating for the supremacy of any of these three models — they can all be effective in different circumstances. Just as we need teams, star players and effective coaches, we need various tools to control, steward and leverage data in a way that best serves us as individuals and society at-large.

In future posts we’ll discuss which of the models offer the most promise under certain circumstances and contexts.

— — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Further reading

Check out these resources if you want to dive deeper into data stewardship:

Ada Lovelace Institute & UK AI Council: Exploring legal mechanisms for data stewardship

Aapti Institute: Data stewardship — a taxonomy

Mozilla Foundation: Data for empowerment

 

Share on facebook
Share on twitter
Share on linkedin
Share on email

RELATED POSTS

Connecting public voices with public policies

Identifying effective means of increasing citizen participation in decision making is timely and relevant considering the impact of the pandemic on populations and governments preparing […]

A recipe for a good shopping

Public procurement can be positively influenced by open contracting. This type of tendering enables all the interested individuals and legal entities to easily access data […]