How can technology support the future of public-interest journalism?

On 24 June, Sourcefabric  convened a panel of four media experts for a webinar on how technology can support the future of accountability journalism, which we define as news that speaks truth to power. Representing the fields of journalism education and training, investigative journalism, and public policy, panelists Mindy McAdams, Paul Radu, Mira Milosevic and Adam Thomas contributed insights and practical examples from their work on what is happening in journalism today, as well as their hopes for the future.

In particular, they highlighted the ways that journalists have begun  working together to solve their own problems. These collaborations fall into four broad categories:

• Project-based: a group of news organisations working together for a specific investigation, like The Panama Papers.
• Purpose-based: the Organized Crime and Corruption Reporting Project (OCCRP) is an example of journalists forming a network to investigate the cross-border nature of organised crime.
• Policy-based: partnering to advocate for journalism in government or multilateral circles. The Global Forum for Media Development (GFMD) established the Dynamic Coalition on the Sustainability of Journalism and News Media (DC-Sustainability) to represent the interests of journalists and media organisations within the UN Internet Governance Forum.
• Crisis-based: working collectively to weather financial or political pressure. For example, the European Journalism Centre (EJC) launched two recent initiatives to support journalists during the coronavirus pandemic. The COVID-19 Collaboration Wire, set up with Hostwriter, aims to provide a central point for commissioning stories related to the health crisis, while the Freelance Journalism Assembly is a platform meant to connect and support freelance journalists over the longer term.

At the same time, our panelists also identified specific cultural shifts and changes in mindset where journalists still have more work to do, including:

• Becoming more engaged with the needs and desires of readers.
• Adopting design thinking, computational thinking, and antiracist thinking.
• Gathering data and information proactively and independent of official government sources, such as operating their own sensors to monitor environmental changes like water quality.
• Extracting data from large unstructured repositories (such as printed documents). 

Technology is the key enabler for all of these efforts; it is the glue that holds together current collaborations, and is the catalyst that can spark nascent projects. Panelists pointed to specific examples, including:

• International databases such as OCCRP Aleph
• Data visualisation software
• Scrapers and sensors  
• AI to process large datasets, as well as specific programming languages like SQL and Python
• Secure communication tools like Signal; dedicated server farms
• Digital spaces for journalists to come together to share expertise, ask questions, and build communities of practice (including Slack channels, listservs and member-based websites)

The full transcript of the webinar follows below. Click here to watch the recording on YouTube.

Transcript: How Technology is Shaping the Future of Accountability Journalism

Let's start with Paul. Could you tell us a little bit about the journalistic needs that drove the creation of a tool like the Investigative Dashboard , which your co-founder Drew Sullivan calls a research desk for the world. Specifically, does a tool like this allow journalists with little to no technical skills to do data journalism, or is a certain amount of data literacy a prerequisite to use Dashboard? 

07:46 Paul Radu. The Investigative dashboard and all the tools that we developed at the Organized Crime and Corruption Reporting Project are tech tools designed around the investigative process. [In the past] there was hype around hackathons, for instance, but at some point, we turned that concept on its head and created “investigathons,” and that's the concept behind what we're doing at OCCRP. We have journalism driving the tech, not the other way around, and I think this is what has helped us build a successful organisation. It all started more than 10 years ago now, when we, a group of journalists in Eastern Europe, realised that we were investigating the same thing over and over again. You know, corruption, in all its shapes and forms, it's always the same. You have the high-powered people stealing money from the masses. And there's things in between – offshore companies, lawyers, accountants, and hackers sometimes helping these politicians and organised crime to steal the money.

09:03 Paul Radu. So, a bit more than 10 years ago, I was in Sierra Leone, in Freetown, working with the local journalists. They helped me a lot to get to the bottom of a story in which oligarchs from Eastern Europe were investing in the mining industry in Sierra Leone. During the course of our work together, this journalist asked me, ‘So, I'm helping you with this, but, you know, I need some help, too, because your one instance of one oligarch from Eastern Europe coming into Sierra Leone to exploit minerals is just one example. There are many oligarchs from other places in the world that are doing the same thing. Can you help me get down to the bottom of those stories?’ This is how the idea for the Investigative Dashboard was born.

I'm very proud of our work. But it's limited because we have limited resources and we're actually not that many (people). This universal dashboard was our initial attempt to assist the investigative reporters of the world and the activists interested in investigating matters that were of concern, to better serve the public. The dashboard was born out of journalistic need and out of the realisation that what we're doing as journalists is very little and there should be a lot more of it.

11:26 Paul Radu. I can just add a few words about the initial structure of the Investigative Dashboard, because that's the point from when we started all our other tech tools that are much advanced now compared to when we started 10 years ago.

There are three axes where we try to develop: One was assisting people with research, and namely research on companies and other types of [organisations] for which information was hard to get. The other one was to index as much data as possible to create a database that's meaningful enough for the world's journalists. And the third reason was to try to make this information more digestible to people. That was where we created these tools to visualise information. 

All this was built on the backs of investigative mind power. All our investigative reporters engaged with the Investigative Dashboard at first were people with a lot of experience, with a lot of context in their heads. And they were willing to put all this together in order to help other journalists elsewhere and activists be able to break really big stories that were of huge public interest.

You said that people are coming to this with a lot of contextual knowledge. Are they also bringing technical skills as well, or is it really something that is meant to instantly crunch data and visualise it for people?

13:07 Paul Radu. A lot of the people who were initially involved with the investigative dashboard were not tech people. But as we went on, we realised that we needed skilled people to be able to further develop these tools. And right now, I think the mindset within OCCRP is that most investigative reporters that work with us have a good level of knowledge, but at the same time, they realise that they need the help of professionals. Our developer right now, when it comes to our data services [Friedrich Lindenberg, OCCRP senior data editor], and many others. So, it's this combination. But always on the back of these tools, data needs to be useful to investigative processes.

That's a good segue to go back to Mindy. What do you think are the minimum skills of data literacy for young reporters today? Where do you start with students?

14:33 Mindy McAdams. With students, we feel like there is a minimum amount that all journalism students should learn. They should know basic descriptive statistics. They should be able to just identify, to be able to have a reaction to a study or a survey or a poll to say, ‘I don't think these data look sufficient or look like they were analysed correctly.’ So, that's the thing that hasn't always been taught to journalists, but I think that's now basic. They should be able to work with data that's in rows and columns, whether that's a spreadsheet or database, depending on their level or what they're trying to do. They should be able to work with people who can do data analysis, even if they don't know enough to perform data analysis themselves.

Not every journalist needs to be able to code to the level where they could produce something with code. But I think it's really helpful if journalists have had a little experience with code, like even if it was a high school class or something like that, so that they understand how much it's not magic, and how much it is possible for people to fail. For example, when you're seeing some kind of data analysis, you want to be able to ask the questions or find somebody to help you ask the questions, ‘Is this valid? Was this conducted in a proper data science way? Or, are these numbers some kind of spin that a political figure or a government official is putting on these data? That's where we are now. It has a lot to do with the data themselves, rather than [needing] to know this tool or this technology or how to write Python. 

The other skill you teach in your courses is building applications that scrape data and put things together. Why is reassembling data important for accountability journalists?

17:42 Mindy McAdams. A lot of times the data that journalists assemble and offer to the public with some kind of interface, like a searchable database, were not available before. So, sometimes the journalists obtain the data through Freedom of Information requests. But they might need to scrape it, either because it's been denied to them, or just because they're told, ‘Oh, this is on our website, but it's not in a very usable form.’ The journalist with a certain skill set will say, 'Well, if I scrape that data, put it into a database, or even just, you know, have a JSON file, which is JavaScript, I can make it available to the public in a really easy way.’

An example of that is ProPublica, which did a project at the end of last year called Credibly Accused. It's about Catholic priests who's been accused of some kind of sexual abuse. And what they found was something like 178 separate lists existed of these priests. Now, the Catholic Church as a global organisation never made one big list, but in the United States, there were 178 separate lists. And so ProPublica collected all the lists, put them together in one database, and if you want to search that database, because you or a relative might have been affected, there's just one search box. In that box you can type the name of the priest, the name of the diocese, or the city. You can type one of those three things in one box, hit a search button, and see if that priest or that diocese is in the 178 lists. If you find a particular priest and want to look at his record, you can click that, and you can get all the data that's available about him from a particular list. That information was not available to the public at all before. And of course, what it allows is, if you wonder if the priest who you're familiar with now was accused somewhere else in the past, was he one of these priests who were moved from one church to another, you can now find that out for yourself. That's a great public service. It's something that neither the Church nor the government nor the police were providing to anyone. And, the important thing is, it wasn't available in a usable form before because the 178 separate lists were all different from each other, they had different formatting.

That's a really great example. Going back to Paul, it sounds similar to a lot of the work that you have to do kind of piecing things together and information that's fragmented and hidden in some way. In fact, at OCCRP, your motto is 'It takes a network to fight a network.' I wondered what you have learned from following organised crime syndicates? How has that informed the tech that you use to do journalism, especially in terms of your communications and your security?

21:56 Paul Radu. We're in constant touch with organised crime. We do have to talk to these people. What's really worrying when you cover organised crime and corruption is that you realise the extent of it. And you realise how tech savvy these people are. I say they're a bit ahead of us [journalists] when it comes to technology. I'd say they're about maybe 10 years ahead of us when it comes to using the technology employing technological setups that are very, very costly. A few years back, we looked in Eastern Europe at the ownership of ISPs, internet service providers, and one of the conclusions of that investigation, which covered I think 12 countries, was that there's huge infiltration of organised crime in the ISP ownership. Now, that's continued, and unfortunately, what we're seeing is criminals and large criminal groups managing and organising their own server farms, offering hosting services to people that are not even aware that they're using the services of these criminals. Sometimes it's for simple reasons, like they offer possibilities to pay in Bitcoin and people just jump at that opportunity. And that's fine, except the problem is when you don't check who's offering you the cheap hosting, and who's promising you this huge uptime and all that kind of stuff. What we're actually seeing is server farms hosting for airports, for banks, for ordinary people and their blogs. But the server farms are owned by criminals who are collecting this data.

What we're seeing right now is a collection era for organised crime. I mean, these people are collecting so much data. I don't think they know yet what to do with all of these data. There are efforts here and there to mine for kompromat, and we're seeing them hosting all sorts of forums where the discussions are between Nazis sometimes, or between other people with interests outside of the public interest, and so on. When you start communicating with criminals, you also see how much they know about secure comms. They're using very, very sophisticated tools and setups. They're using things like Signal, which everybody's using, but they go beyond that with server structure. And so I think they're very, very sophisticated and I think that’s going to be a very interesting point to investigate [further]. Organised crime groups moved from drug trafficking because they had so much money at their disposal into exploiting digital information – your digital information. People are not that much focused on these criminals, but criminals are a lot more advanced I think than journalists. And the methods that they're employing are scary.

Let’s shift gears a little bit. When we think of technologies that can support public interest journalism, they fall into two categories. One is technologies of corroboration. And another is technologies of collaboration. Mira, can you speak about journalism collaborations emerging or interesting projects that are coming together, and what kinds of technologies are supporting them?

27:02 Mira Milosevic. Thank you. And thank you, Paul and Mindy, that was really interesting and is in line with our thinking – that technology does influence how the future of accountability journalism will be shaped. But we also need to think of ways how we can influence those technologies, especially at the level of architecture, decision making, values, and principles, and how these will be built in the future. How can we, as members of a journalism and media community, impact that to actually be a space that will work for accountability journalism and for accountability to all citizens.

What we [at GFMD] do is try to create these spaces within some international policy and regional and national decision making platforms, so that we can come with the knowledge that Paul has just presented, with the knowledge that we have from Europe from Adam and his [EJC] partners, for you and your partners at Sourcefabric, and go to these decision-making circles such as IGF [Internet Governance Forum], and try to bring the perspective of journalism and how journalism can survive and thrive hopefully, in these digital spaces.

There are two areas where we focus. First is to create the spaces and platforms. The second is to facilitate cooperation and collaboration along the lines of needing to fight networks with networks. A couple of things that we are doing at the moment. We have a dynamic coalition in the Internet Governance Forum, where we bring journalists from Latin America, from Africa, from Asia, and Eastern Europe, but also from the US and Western Europe to [discuss common challenges]. What are the issues that they're facing with security? What are those mechanisms that allow organised crime to own these IP addresses? But also to see what are the aspects of financial viability that you mentioned, what are the structures that are allowing or more often not allowing accountability journalism to be valued and to have its own kind of sustainability online.

One of the things that we're doing at the moment is looking at the Digital Services Act in Europe, and the consultation processes in place. From the perspective of accountability, one of the questions in the 60-page consultation survey is what information should competent authorities make publicly available about their surveillance and enforcement? These are the things that we try to look at from the perspective of journalism. We're also looking at the Santa Clara Principles, self-regulatory documents from big platforms about content takedowns. For instance, if Paul is working on a story or Mindy's students are reporting something for three months and they use a lot of resources to produce it, there needs to be a certain procedure that will be followed if someone wants to take that content down. This is another area where collaboration is really important.

From the tech perspective, we are looking to provide spaces for a community of practice to connect our members – like Sourcefabric, the European Journalism Centre, and OCCRP – to share knowledge about these tools, and to provide journalism organisations and journalists with knowledge so that they don't have to go to the most popular, the most commercial products that are not respecting maybe in all cases the human rights principles, privacy, and are not transparent about their terms of service. We also believe in the power of collective action. We have launched a call for support to journalism in the context of COVID, together with around 180 organisations. And finally, we think it's a breaking point in time. We think it's the right time for us to be talking about new ethical principles of accountability and journalism and information in online spaces.

Some of the things that OCCRP is doing are fantastic. For instance, a collaboration is happening at the moment with OCCRP, GIJN, and an international consortium to support what Maria Ressa and Rappler are doing in the Philippines. That has been facilitated and enabled by technology they're all using. 

Additionally, there is a collaboration going on at the moment in Brazil, where local journalism organisations are collecting data on COVID infections, because the state has stopped publishing this data, and investigative journalists and accountability journalism is helping this. There are a lot of initiatives that our members are implementing at the moment. I can talk about that a bit later.

Following up on collaborations, maybe we can bring Adam into the conversation. Recently the EJC launched a couple of initiatives responding to the current health crisis, the COVID-19 Collaboration Wire launched with Hostwriter, and the Freelance Journalism Assembly. Both are new programmes, so it's very early days. But how do you see them evolving, perhaps into longer term support for the media?

34:25. Adam Thomas. Both of these projects actually had a life long before COVID-19, and were somehow evolved very, very quickly. I think it shows that the journalism and the organisations working within it at the minute are able to rapidly evolve things and listen to the needs of users to be able to develop new software and new techniques together. Both of those projects in particular, what they indicate is actually a real change in how we're thinking about how we design programs [at EJC]. I think a lot about communities of practice right now. We are at a really dangerous point for journalism in terms of its sustainability and its resilience. And so, one of the ways forward, I think, is to empower communities of practice, especially with journalists to be able to start to solve these problems themselves and come together. I see the European Journalism Centre as a site for doing a lot of that work. And I think we're seeing very exciting things happen in that regard. We had 900 freelance journalists sign up to our [freelancer] programme in the last two or three weeks. We launched our first webinar yesterday [23 June 2020], which was more of an interactive workshop with 100 people. We're just seeing a real need, like journalists themselves feel like they have to get together and start to fix some of their own problems, because they can't sit around and wait for other people to do it.

That's exactly the kind of spirit that Paul and other successful teams have shown. You have to find a way or make one, to quote a Sourcefabric phrase from the past. I think that's the first thing: communities of practices are a really important thing. Second, of course, is collaboration. As information becomes more globalised, and journalism becomes more globalised, some of the traditional boundaries between users that are inherently competitive start to break down. I think we could start to see more open collaborations and investigations, like the type Paul runs, or the Panama Papers. These are reliant on major international collaborations of journalists that, first, weren't possible because of the technology [in the past], but second, perhaps because of mindset. And I think that's one of the strongest areas that we start to see things change. These aren't necessarily technological changes, but they're definitely mindset changes …[and] some of them are born of a product thinking, the idea of rapidly testing and prototyping a product, which has come into newsrooms in the last 10 years or so.

This is about computational thinking, and journalists getting much better at being able to think, ‘How can we use computers to solve problems?’ That type of thinking should be taught in all universities. As Mindy mentioned, you don't have to code to know how a computer would solve this problem or work with someone who does, I think we're seeing a number of shifts in the mindset of education as well, and that really complements the type of technology. As the bad guys get more sophisticated and the data sets get bigger, we don’t only need improvements on the technological side. We need them on the collaboration and thinking side, too.

Adam, you mentioned communities of practice. Can you talk a bit more about that, as it may not be a term everyone knows?

38:50. Adam Thomas. Yeah, sure. Thank you for jargon busting, I should have explained that. It’s just when communities who are trying to solve a problem together or have a common interest come together to solve that problem themselves, and then think about sharing the learning within that group and then more broadly. It's nothing hugely ground-breaking; it’s a term that goes back to the 50s and 60s. But it's really an effective way of learning. And I think it becomes more and more important in this current environment, to think about that as a form of education. Codevelopment and collaboration are core parts of journalism education, but also [essential] to the future of the industry going forward.

We haven't spoken about the public too much here, but that's obviously key. We spend a lot of time as an industry looking inward and we have to change a lot. I don't think we're thinking enough about the public. And what does the public need to know? What does the public want to know? And when we think about a ‘community of practice,’ how can we develop a community of practice that also involves audiences, so that we're reporting with those audiences for them, not only on them. This comes back to the principles of engaged journalism for me, which is really asking citizens what they want to know and involving them in that process. This can actually be crowdsourcing, but it can also be really engaging people for tips for story ideas, for things that they want to understand about their politicians and what they want [their leaders] to be held accountable for. But there's several things that journalists have to change before they do that.

Black Lives Matter is really important, this sharpening of focus. We need to think a lot more about diversity inside of newsrooms. You know, if you look at some of the data coming out of Germany at the minute, 25% of people in Germany have some kind of immigration background, but only 2% to 5% of journalists share that background. So, we have fundamental problems inside newsrooms that we have to fix, and that if we don't make our newsrooms more representative, we're always going to be failing to represent and inform the public in the fullest way possible. 

It's about the mindset of thinking of our public or thinking about our audience as part of that community of practice, who we can work with. That extends to the reporting, but then it also extends to how we distribute and how we engage people in that news. One thing that we have seen from COVID-19 is news organisations have gotten much better with community calls, virtual town halls, running WhatsApp groups, all of these types of things to really try and engage people more deeply in the work that they do. Long term, that's where the survivability of the media industry is going to come from, because that's going to be what helps justify the value of journalism and helps people pay for journalism optimally. 

Mindy, how about your students? You're dealing with digital natives and your students have grown up already in a different world. Do you find that you need to teach them how to digitally engage with audiences, or do they just come in knowing how to engage? 

42:17. Mindy McAdams. It varies a lot among students. People who don't teach in universities and journalism schools don't always realise that not all students who sign up to study journalism actually want to be any kind of traditional journalist. Sometimes they're not comfortable going out and interviewing people face to face. They're not comfortable cold calling people on the phone. Not every student who's chosen journalism has really chosen to be a traditional reporter. This is why, you know, I agree with everything Adam was saying. And also the idea that we can’t only report what we think communities need to know, but we need to report more with them as our collaborators. That will open up other kinds of ideas, because the people in the community aren't always comfortable with us. And so this is a thing we see too with students, the students go out and they're doing this for the first time and they're uncomfortable going into a community they've never been in. I'm not even talking about like racial and cultural differences as much as the student is from another part of the state of Florida, and they're here in our county, and it might be very different from where they came from. And they go into a community that they are unfamiliar with, they're uncomfortable, and the people talking with them are uncomfortable. And sometimes the people talking with a student are uncomfortable because the student looks so young. So, the people in the community don't really feel trust for them. 

These are the kinds of issues that we talk about at journalism conferences, but I feel like there needs to be a lot more intermingling between communities and journalists and talking about these feelings that people have. When they're trying to talk to strangers, when the community members are trying to put trust in a stranger who they may fear will misquote them or misrepresent their concerns, or even them as people, who might feel like their community will be misrepresented because maybe it has been in the past.

One other thing I want to address related to what Adam said. Data journalists in particular often see themselves as a community of people who have this certain skill set, and they are very willing to help each other, very willing to even do work for others, provide resources. They might say, ‘Send me your data and I'll see if I can help you with that problem you described.’ The EJC is a great organisation and has great training. Another is the Investigative Reporters and Editors, which has Slack channels for data journalists, and a listserv for everybody, and people are really willing to help. They ask questions of each other: ‘Have you ever worked with a data set in this format? How do I work with it? I've got this giant collection of PDFs that I have to scan and analyse, what's the best way to do that? I never did it before.’ People are really generous with their time and their knowledge, and that is even across organisations. For most of my life, newspapers in particular in the US have seen one another as the competition, with the whole idea that they are competition. Journalists felt ‘I shouldn't share with them, I should scoop them, I should beat them, I should have the story first, I should keep my information secret from them so they don't steal my story.’ But that [mindset] has changed a lot just in the past several years, particularly as we've lost a lot of jobs in this industry and so forth. But I feel like more journalists, and particularly younger journalists, are more open to sharing knowledge and even sharing data and skills with people who work for another organisation, whether it's in the same town, the same city, or in a separate newspaper chain or news organisation. Journalists at these giant conglomerates might think ‘You work for this one, and I work for that one, but, behind the scenes, we're going to help each other analyse this data, and maybe even share the data and each of us will run it in our own separate publications or broadcast it.’

That's really interesting and hopeful. Maybe we could come back to Mira. You mentioned that GFMD is working with some groups in particular. Could give us more examples of these collaborations and knowledge sharing opportunities?

48:26. Mira Milosevic. We’ve created a number of groups and participate in groups that discuss all these issues. One that we run is this Internet Governance Forum, where we bring people together to report on all these issues. For instance, one thing that is running at the moment is the International Journalism Center, which has tools and resources for reporting on COVID in six languages. That's something that we share within these groups. More importantly for us is on a policy level; one of the groups that participates in our Internet Governance Forum is Free Press from the US, that has just calculated that over the last 15 years, local journalists alone have lost between $850 million to $1.2 billion in wages. They also calculate that local journalism in the US has lost around maybe 20,000 jobs. If we think about the impact of technology and the future of journalism, this is one of the core areas where we bring people together to work towards addressing this.

We estimate that the situation on a global level is five-fold to what we're seeing in the US. So, if we're speaking only about local reporters, local accountability reporters in the traditional sense of the word, we probably have lost during the last 15 to 20 years around 100,000 reporting jobs globally. 

If we are talking about future accountability journalism, this is something that we feel we have to urgently address to find a model that will allow these people to do their jobs and [get paid] beyond only small models that work in certain situations. This community of practice, of sharing what works, is something that you do at Sourcefabric. One of the things that we also feel is important is understanding how journalists receive signals from these audiences. How do you go beyond Google Analytics to know how much your audience engages. But, on the other side, what signals are you sending to big platforms about the quality of your work, and how do platforms and the digital spaces tell the difference between credible accountability journalists and the noise that's dominating conversations. 

Finally, we are working with groups like Media Diversity Institute ...to look at the situation that Adam mentioned in newsrooms, and what are the tools and resources that we can all use to to improve the situation both in terms of representation, diversity reporting, and recognition of the problems that we have, but also in terms of sustainability? How do we keep these jobs? How do we bring in new jobs? And how do we continue to have the ability to report on corruption, to report and hold those in power to account?

Paul, OCCRP is a great example of journalists collaborating amongst themselves to do just that. Can you talk about building trust with the public, and encouraging journalists to work together?

52:50. Paul Radu. While the general situation with journalism is dire, when it comes to investigative reporting, there has never been in the past 100 years a time as good for Investigative reporting as it is now. Just think about all the centres for investigative reporting that have popped up all over the world in the past 10 years. There have never been in any traditional newsrooms as many investigative reporters. Investigative reporting was, in traditional newsrooms, side work for people who were in high regard in the newsroom, but would produce very little from time to time. I mean, excellent articles and all that, but very limited. 

I think we're still limited but better than it used to be. There's a lot more investigative reporting, and there's a lot more focus. The investigative reporter used to be like a grasshopper, jumping from topic to topic, from subject to subject. They might tackle trafficking human beings one day, and some other issue the next – like high school dropouts.That's all good, except I think investigative reporting is really efficient when it builds up, when it builds on previous experiences, when the technology that's built to help investigate reporting is added on top of those experiences. 

I think that's what we're witnessing right now. I'll give you an example. This was a couple of days ago when we talked with all these media organisations around the world to start this initiative to work alongside Maria Ressa and Rappler to continue their work. The discussions in that group, which includes investigative outlets from all over the world, was what tools should be used to exchange data and comb through stuff. Everybody had their own tools. They worked so much in the past 10 years to develop and create their own wikis and their own secure channels and their own data analysis tools. It was amazing. I mean, if you tried this project 10 years ago, you'd be like, ‘What can we use?’ So this is something that changed quite a bit. Of course, this comes with lots of problems as well. 

You're also right to address the issue of public trust, and I think the main problem right now is that we're offering to the public very good quality content that tackles issues that are of public interest, [but perhaps is still limited quantities]. What we're seeing for instance at OCCRP is that the public comes to us and says, ‘Oh, but you should also deal with this problem and with this problem with this problem.’ Lots of those problems are an iteration of what you already offered as context, as investigative reporting, except you can't tackle all those problems. And then the public gets disengaged from your effort, because you didn't help them on their specific issue. This is where we need to work right now, to realise investigative reporting is great, but it's limited. We need more people to be able to use these tools, to use to be able to use content. We need a lot more technology to be able to develop patterns, because what we investigated just now, like in Kenya, for instance, is the same that you can investigate in Peru, or in Romania, my own country, or elsewhere. I think that's growing and that's changing. With this big collapse across borders, journalists are becoming more humble, and that means they're becoming more prone to cooperate across borders. In fact, we're seeing high level newsrooms cooperating now with smaller initiatives in countries that were not even considered before. This is really great actually, when it comes to investigative reporting. 

We have time for one more question for everyone, which is: Is there something specific that gives you hope right now for the future of accountability journalism? Paul, would that be the increased sense of humility and better tech that you just mentioned? 

57:29. Paul Radu. Yeah, one would be humility and cooperation based on the recognition that we need to serve the public better, and we need to give the public the right tools for them to do the investigation. 

The other is that we're relying so much on government generated data, but we need to take action [on independent data sources]. I'm seeing that to some extent, for example, with people moving a bit more into ‘sensor journalism,’ like planting their own sensors to track down flights, or having sensors to see water quality. These sorts of efforts are very interesting because they don't have to rely on government sources, especially when we're working across borders.

58:36. Adam Thomas. I think if you look at COVID-19, and the reporting that's been around that recently, there's been so much incredible reporting, whether you're looking at something like the Washington Post’s data visualisation, or some of the really deep data work has been going on at the Financial Times, which is taking what is pretty complicated data that could be represented in multiple ways and making that accessible for audiences outside of a paywall. So, the reporting there has been amazing, the data and the collaborations have been great. But also, it's had some really deep impacts. In the UK, where I'm from, it has fundamentally changed government policy and held them to account. I think we've seen journalism fulfilling its mission in such an intense way in COVID-19. And I think we've seen public recognition of that as well. If you look at the Digital News Report from Reuters, you will see that some of that trust starts to come back in certain areas, which is something we should hold on to and build on.

59:37. Mindy McAdams.  

I wanted to throw in the idea of artificial intelligence, specifically, machine learning, as a double-edged sword. [For instance,] scanning PDFs. It may be government data, it may be corporate data, it may be health data, but no matter where it’s collected, it hasn't been prepackaged. You've got way too much stuff for humans to deal with, particularly [given] the reduced size of newsrooms. So, journalists can use technology to scan the PDFs and make them human readable, or use optical character recognition, which means things can be selected from it. 

This was done with the Atlanta Journal Constitution's examination of sexual abuse by doctors of patients. Humans figured out some key terms that would appear in complaints against doctors and said, 'Okay, so go through this giant amount of scanned documents and pinpoint for us humans, which ones we should look at, which ones look likely to contain a case we should investigate.' In that way, you can go through hundreds of thousands of documents, but of course, certain expertise and so forth is required. Mistakes can be made but the example that I just gave is one where what you're really asking for is it to cut through the ridiculously large quantity and show me documents that I should pay closer attention to and I, as the human, can decide, ‘Okay, no, that one wasn't really worth looking [at], or, Ah, this is the stack that we should investigate.’ We've got tools now that were never available to us before for all the same reasons that business and so forth are looking at artificial intelligence systems. We've had changes in technology that have made these systems more usable, more practical. Fifty years ago, it was like crazy stuff people were doing with AI. Now, there's very, very practical, practical stuff that journalists, researchers, and even social scientists are using. 

But the bad side of the sword is that many mistakes can be made. And I think everybody should be paying attention to everything that's being written about facial recognition, because that is one of the most problematic areas right now – the use of facial recognition technology to identify people as some kind of threat or wrongdoer. And they're being wrongly identified, because a system that's made of code can be flawed, and many of them are. We have to be knowledgeable and ethical about how we employ these systems. And we have to be transparent as journalists about how we're using them. 

That’s my last point, there's accountability journalism where we are holding the powerful to account. But there is also journalistic accountability, where we hold ourselves to account. And we need to be transparent about how we look at the data, what tools we used, and what possible problems there might be in our analysis.

64:42. Mira Milosevic. My biggest hope is that those communities of practice and those networks like OCCRP are growing and there are more of them. Those are the only sustainable solutions for us to address both the needs of our citizens and audiences, but also the big policy decisions and questions that are facing us in the future. To close, I will quote Maria Ressa, whom we mentioned a couple of times today: 'Know that no matter how much of a superstar you are, you cannot accomplish anything meaningful alone. Build and strengthen your community. Rappler has built communities of action. I just hope it's enough to protect our democracy.'