Is it really Tech/Data/AI for good?

The term “X for Good” has been bothering me for a while. I too have used the phrase tech/data/AI for good before. However, it has become a catch phrase for too many things, but mainly, a feel good phrase (see what I did there?) for tech companies thinking they can compensate for some dodgy stuff they are doing by working with nonprofits. I was very excited to see this taxonomy by Jake Porway trying to define this field. However, the more I read it the more I got upset and worried that funders will take this taxonomy and suggested activities as is, and that was ALARMING to me. This taxonomy doesn’t answer key questions— Good for whom? What power dynamics exist? And who gets to decide what is good? This blog will try to untangle these questions and offer some alternatives.

First of all — What is good?

The methodology of the taxonomy defines good as “reducing harm” and “increasing benefits from”. However, both of these are very subjective. As a philosophical exercise — if I am a pro-life activist that wants to track abortion clinics to reduce harm for fetuses, is this a data for good initiative? For some people, it is reducing harm and increasing benefits from data, but for others it isn’t. If I am a company tracking millions of people and ignoring climate change but I give free storage for nonprofits, is this enough to say that I reduce harm and increase the benefits of data?

What the taxonomy piece is lacking is something that the data feminism framework keeps advocating — examining power. “X for good” initiatives should actually be framed as initiatives that are promoting justice and disrupting power dynamics. Justice is not only limited to the rule of law — How do we bring justice to our climate and ecological systems? How do we create health justice by eradicating cancer for all?

You can say, “Mor, justice is a subjective term as well.” True, philosophers have been arguing on the meaning of the term for centuries. However, justice takes the discussion a step further than good. Something can be done for good, but it’s not always fair or distributed equally. Justice makes sure we discuss what is fair and how we are doing things. As a good friend of mine told me — “Good should support justice but most times in these type of initiatives it supports tech solutionism”

The term justice is not something the tech industry finds easy to deal with. The traditional tech world is filled with stories of startups and unicorns that create tech/data/AI solutions that are sold for millions of dollars. So let’s acknowledge that the tech world is about fiscal profit. The world in which nonprofits, government and academia are on the opposite end, they usually don’t create any fiscal profit. Therefore, investment in tech/data/AI within these institutions doesn’t (usually) come from the market. This is where the invisible hand fails over and over again. This is where I see most “X for good” initiatives coming in — charity and good will of others to donate to nonprofit on their terms. It is actually a big problem.

But Mor, it’s important to donate time, resources and money to those in need! Sure it is, but this also creates power dynamics and silos, which we need to acknowledge. Why should social issues not be dealt with in cooperation? My experience shows that usually when it comes to X for good initiative the solution is what the techies want to work on, not necessarily what the nonprofits need. Like pushing AI instead of looking at data basics, or pitching NFT because they are the latest cool thing but not thinking about the environmental toll of cryptocurrencies. It works to a point, but it doesn’t help with the important and boring stuff like stability and long term data sustainability. The framework of Build With Not For was created for civic tech work eight years ago, we should refer back to it in our data work now too.

So I suggest we should change the term “X for good”. We should be bold and honest about what we are doing here — we have data for (fiscal) profit and not for profit data. I think that data for non-profit creates long term profit, but this is another blog post.

What about humans?

The second issue with this taxonomy is that it’s very technocratic. It misses the human aspect of data. When the suggested data pipeline discusses humans, it refers to those who are the data people, the techies. There is little to no mention of other roles. The world does not and should not revolve around data. Data without context is hazardous.

What can we add to this taxonomy? For starters, if this taxonomy is going to be used, it needs to ​​be written with the assumption that people are before data/tech. All the people involved must be responsible and ethical with data and tech.

In addition it misses terms culture, team work, and emotions. Emotions such as fear of change or anxiety from not knowing where your income to cover a data person will come from. The data pipeline starts from problem-solving when actually, the biggest challenge I have encountered as a data person working in nonprofits is that people are afraid or don’t know how to ask questions. We retreat to solutionasing than questioning and exploring because it is easier for us not to deal with hard emotions that questions bring.

Another example is that the taxonomy focuses on communities for data people but not on communities where everyone, even those who are not data people like the marketing or HR people are welcome to come and discuss data. Data is political and in order to make sense of it we need to have other people in the room, even if they don’t know what a pivot table is.

Interestingly enough, data.org released a couple of resources that touch a bit on these issues, like the data for impact or the resource library and data maturity assessment, but the taxonomy seems to be a separate line of thinking from them. I also think that those resources can address power dynamics even more.

Not everything that is given for free to charity is X for good

The taxonomy defines storage and off the shelf software as “for good”. While I am thankful for the ability to get free of charge storage or software as a nonprofit, I don’t think this should be considered as X for good.

Firstly, a lot of these schemes are limited by space or features. For example, big charities still need to pay for some of these tech solutions since they go over the limit of users/space. This is not for good, this is a business strategy — let’s hook organisations to our service and then ask money for money if they need an upgrade or more features. Fair enough, companies are trying at the end of the day to make money. It means we can’t call this strategy “for good”.

Secondly, the fact that some companies have “for good” initiatives doesn’t mean they are overall doing no harm. Sometimes, those free schemes for non-profits look like a bit of a cover up to other harmful stuff that they do. So the fact that Google is offering free stuff while at the same time having scandals in its AI and ethics team creates a bit of a conflict here. It leads us to the question of power again — what is the power dynamics of big tech companies and nonprofits?

Lastly, the taxonomy suggests the following “If you’re a funder who believes we need more storage available in the social sector, you might fund them to provide it”. I hope I read it wrong, but I don’t see why Amazon or Google, multi billion dollar companies (with a turnover bigger than some countries GDP) need to be funded by other funders specifically to help the social sector? Why can’t they give if the cause is really important? Just putting it out there.

OK, so what should we do now?

  1. If this taxnomy is a key resource for data.org, then it needs auditing with a power dynamics lens on it. Data.org can’t claim to democratise data use without this lens. This taxonomy needs a second iteration and needs to be shaped by different communities. If needed, pay organisations to give this feedback. Value their opinion and work. Don’t put the burden on the community to give you feedback, like I did it here. Not everyone has the privilege to write a lengthy response like I did.
  2. Stop speaking about X for good, start speaking about what is a non-for-profit data and tech. Start speaking about justice. What does it mean that vital data infrastructure doesn’t get created because it will shift power dynamics or won’t create revenues.
  3. Stop the silos. We need to work together on a topic, share expertise and experiences to create better solutions. Siloing problems because they are “for good” is creating a broken ecosystem where we fund short term solutions and not actual culture change.
  4. If you are a funder, consider looking at core funding for data and tech work, instead of project funding. If we want to break the silos we need to make sure that we not only train people, but that nonprofits have the budget to actually employ them in the long run. It is also important to consider investing internally as well as upskilling your own team on data strategy and how data can improve grantmaking and grantees as well.

Some related resources

(If you have ideas for more comment on this blog and I will add them to this list)

Humanitarian Digital Ethics: A Foresight and Decolonial Governance Approach

Civictech.guide

Decolonizing Digital Ecosystems

Decolonising data — Data4Good Fest

Big data for climate action or climate action for big data?

Many thanks to Ana Brandusescu, Heather Leson, Wakini Njogu and Tom Watson for their comments, input and ideas for this blog.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Mor Rubinstein

Mor Rubinstein

Data strategy. Feminist. Politics and history geek.