Truths and half-truths in Oslo: the technological keys to fact checking

Montserrat Rigall, journalist from RTVE, transfers the main lines of learning extracted from its presence at the world meeting of fact-checkers and advances how RTVE's IVERES project proposes to combine and create tools based on AI.

Hundreds of professionals members of fact-checking entities and organizations from around the world gathered in Oslo last June. The first in-person conference in three years has served to show that misinformation inundates us more and more on a global level every day and that its forms of delivery and consumption are increasingly diverse. Furthermore, emphasis has been placed on the negative consequences for a democratic society that disinformation and post-truth can have by fueling populist discourses.

Different panelists have emphasized the difference between misinformation and disinformation. The misinformation It is false information that is shared on social networks, the media, telephone lines or interpersonally even though there is no intention to deceive. The disinformation, on the other hand, is false information created or shared with the deliberate intent to deceive.

According to John Donovan, a researcher specialized in disinformation in the mass media, the manipulation of the media serves to “derail democracy” in our society and to avoid this we must “train journalists and citizens so that they learn to discredit” certain information malicious

---

Oslo, capital of verification

Checks and verified information have filled the Norwegian capital with truths and half-truths for 4 days in June 2022. The summit has shared knowledge from the academic world and the practical world of verification against the threat of disinformation. There is no doubt that there are great disinformers, some with political interests, others with economic interests, and all of them execute waves of contaminating disinformation campaigns with relative ease and with little cost, at an almost ridiculous price compared to the damage they cause. Hence, the opportunity to become part of this group is of special interest for innovative projects such as IVERES, a project financed by the Ministry of Science and Technology with Next Generation funds from the European Union and led by the Spanish Radio Television Corporation and UAB whose nature is to be at the service of the citizen and, according to Law 17/2006, “guarantee the objective, truthful and plural information, that must fully adjust to the criteria of professional independence and the political, social and ideological pluralism present in our society.”

Oslo has highlighted that the identify misinformation It is, more than a social need, a global urgency.

In this sense, it is vitally important to have tools and adapt them to the needs of journalists to help us in this arduous task of verifying and thus improving our usual routines in the newsrooms. Precisely, the project IVERES works to generate an “almost custom-made” tool that is available to journalists in terms of verification of video, photo and voice images, as well as metadata.

Oslo has highlighted that identifying disinformation is, more than a social need, a global urgency. An intense debate is on the table about how and who should prevent the spread of false information that leads to destabilizing the foundations of democracy with institutional, political and social disaffection of citizens.

In short, our presence in Norway makes absolute sense of this project of the Chair of RTVE-Autonomous University of Barcelona, which also has the participation of the Carlos III University, the University of Granada and the Polytechnic University of Catalonia. Currently, in the post-pandemic era, we cannot ignore an international debate of such magnitude that it allows us to add knowledge with those who lead this journalistic subgenre and, in the future, the possibility of contributing our own initiatives.

---

Global Fact and the IFCN

The summit, called Global Fact 9, is the most important meeting of checkers that currently exists in the world. It is organized by International Fact Checking Network (IFCN), born in 2015 in the North American state of Florida and whose headquarters are in the Poynter Institute, an entity financed by Bill & Melinda Gates Foundation, the Open Society Foundations y Google, among others.

The global verification event has brought together more than 500 participants from 69 different countries. Beyond the numbers that continue to grow year after year, it is worth highlighting the new journalistic subgenre that was born with the urgent need to verify information, even more so after a global pandemic caused by covid-19 and the War of Ukraine that has left us all in the greatest uncertainty ever experienced in the postmodern era.

The IFCN prepared, one year after its creation, a Code of Principles which establishes that verification has to be transparent (that is, non-partisan) and committed to key fundamentals: equity, transparency of sources, financing and organization, methodology and open and honest corrections. These 5 points govern the work of those assigned to the IFCN which currently bring together around fifty entities around the world, including media outlets, institutions and different companies.

It would not be unreasonable to emphasize that these verifications also have to “be non-partisan” as stated by the Doctor of Journalism, Nereida Carrillo, one of the pioneers in bringing digital verification closer to young people, teachers, families and journalists, through the platform Learn to Check. Carrillo assures, correctly in my opinion, that we must also “verify all political parties.”

Why can it be interesting to be a member of the IFCN? Because it adds credibility to the work of any entity since it is an essential requirement to comply with this Code of Principles. In Spain they are members EFE Verifies, Neutral (startup founded in 2018 by journalist Ana Pastor focused on production for TV and platforms, new narratives and fact-checking), Maldita.es y CHECKED. The presence of technological projects that provide new tools for these tasks, based on Artificial Intelligence, is especially timely and interesting. This is the case of the project IVERES and more specifically the tasks it carries out in terms of technology within the RTVE Chair Carlos III University.

---

IVERES at the IFCN

The ultimate goal of IVERES is to provide a tool that facilitates journalistic work with the help of artificial intelligence applied in the Information Services of the public entity, which is why it pursues, in general terms, the same as the IFCN. Both work on monitoring formats and tracking the impact of verification. The Poynter organization also organizes an annual meeting to promote collaboration between fact checkers, offer in-person and online training, and announce scholarships.

One of the most repeated ideas with which the Global Fact 9 has been that of Baybars Orsek, director of the IFCN, stating that “our collective trust in reliable and authoritative information is under attack by people in power.” Baybars added that “manipulation of the truth makes people vulnerable to bad actors who capitalize on their lack of access to quality information for their own benefits. “Autocratic governments and strongmen around the world are following similar playbooks to censor free speech and dissent under the name of fighting “fake news.”

---

Non-existent protocols

With respect to exchanges of information and agreements with institutions, specific agreements have been carried out to protect electoral periods in Europe with some means but, as of today, there is no connection nor any established protocol. Carlos Hernandez Echevarria, public policy coordinator of the Spanish fact-checker Maldita.es, tells us that his company works with entities to verify some issues on a daily basis or because they are interested in a specific piece of news, but, warning of a risk: “You cannot accept that a public institution imposes on you what is a lie or No".

Sergio Hernandez, responsible of EFE Verifies, assures that misinformation “is very profitable because you achieve a lot of results with very little investment.”

The law is clear about what slander is but, according to Hernández, some institutions skirt it: “Many times institutions carry out initiatives instead of doing the most basic thing, which is being more transparent.” The request for transparency in sources, costs, payments and collections is an idea repeated in Global Fact 9. Sergio Hernandez, responsible of EFE Verifies, assures that misinformation “is very profitable because “you achieve a lot of results with very little investment.”

---

Verification tools

Researchers in Oslo have found 134 data verification tools and services. In the Technology and Fact-checking presentation titled 'Technology to the rescue', the professor from the University of Bergen, Carl-Gustav Lindren, has pointed out: “The approaches vary, but they all seem to share the common goal of providing true and reliable information to a user, whether a journalist, a decision-maker, a businessman, politician or individual.”

From his point of view, he draws attention when he states that “technology is not enough to solve social and political challenges, such as information disorder,” and adds that “in addition to fundamental issues, such as the lack of deep understanding of problems we want to solve, there are also serious technical limitations due to, for example, data problems.” In conclusion, The relief assures that “disinformation is largely a social and political problem that needs a broader approach than technical solutions alone.”

Given the minimum cost What does throwing mean? hoaxes, the immediacy is required.

In this sense, the researcher Laurence Dierickx, states that “algorithmic tools designed specifically for the journalistic investigation process are rarely if ever used,” and adds that “journalists are not aware that the hidden assistant facilitates their investigation process.”

The idea of ​​a fully automated fact-checking platform that can detect a fake in real time, and instantly provide a rating on its accuracy appears to remain challenging, despite some progress. Dietrickx ha evocado algunos autores: “Una herramienta totalmente automatizada que juzgue que una afirmación es verdadera o falsa siempre está limitada en funcionalidad, precisión y comprensión”, según Masood & Aker, 2018.

Given the minimal cost of launching hoaxes, immediacy is required. Faced with possibly false news, the citizen contacts the verification companies through WhatsApp. A chatbot created by the same organizations responds to them based on a database. From there, he investigates himself with his own sources and with the help of artificial intelligence.

---

Most used platforms

Among the most common tools to check the veracity and reliability of images is the so-called initially Inside (In Video Veritas, https://www.invid-project.eu/) that was created within the framework of the European Horizon 2020, and whose current name is WeVerify InVid. This standalone tool or compatible as a Chrome plugin allows you to check the metadata (publication date, recording location, duration, channel where it is hosted, date of creation of this channel...) of a YouTube or social network video to confirm that it has not been created. been edited or manipulated. In addition, it allows you to do a reverse search to reach the source and, among other functionalities, you can enlarge an image in detail through the magnifying glass, recognize characters from a photograph...

Another very common tool among verifiers is the Crowd Tangle that allows you to track links shared on social networks. Journalists can use it to check the impact of a news item, currently only published on Facebook, Instagram, Twitter or Reddit, that is, to know what engagement has had a publication. It has two versions, one free and another from Facebook for the media, because Marc Zuckerberg bought it in 2016, five years after his birth. However, this tool is generating debate as some fact-checkers speculate that Meta will remove it before the European Code of Practice on Disinformation comes into force. When asked about this topic by Meta representatives in Oslo, No one from the company has officially confirmed it..

If what we are looking for is to identify manipulated or false images, we can resort to Assembler, a deepfake detector to discover fake faces. It is a platform that, just by uploading an image, pixels are analyzed to hide details, changes in saturation or brightness.

For when we want to know who captured an image or where its origin is, we have the TinEye, a search engine that has more than 16 billion images attached with two functions in the option TinEye Labs: Multicolr (allows you to find images with the same color) and Extract Colors (allows you to locate colored fonts by simply uploading an image or adding a URL).

The geolocation of Google Maps allows you to identify a scene based on shadows and reliefs. Furthermore, the tool Plus Codes, Based on latitude and longitude, they help define a specific location for a conventional address and do so through a grid system and a set of 20 alphanumeric characters. For example: JJXX+HR8, Seattle. It is very useful if we want to identify the various entrances that a building may have.

Google also offers reverse image search with Google Images and searching for verifications through the Google Fact Check Explorer. It is a tool designed for journalists, researchers and fact-checkers, which is why it is probably little known, but anyone can use it.

---

Interesting agreements

Among the agreements that are being developed is that of IFCN. It will be consolidated in September in the form of the creation of a guarantee seal for those who are affiliated or want to join the organization. It would be very interesting for the project IVERES be in it so as not to lose presence as a means of public communication, to follow the progress of the protagonists in verification and, in a short period of time, to be able to contribute our advances in the matter. After all, the entity RTVE It has an undeniable global reputation, especially in information services, so we can provide them with practical knowledge in the development of journalistic practice.

In another order of agreements, the European Comission promotes a European network of observatories for the study of digital media and against disinformation, led by the European Digital Media Observatory (EDMO). Based in Florence, it encompasses 8 observatories, one of which is Iberifier, an Iberian-Portuguese agreement. With this draft regulation of digital services, DSA (Digital Services Act), the figure of “trusted flaggers” will be created whose job will be to warn of illicit content.

The rest of the members are IN FACT in France, IDMO in Italy, EDMO IRELAND, BENEDMO in the Netherlands and Belgium, CEDMO in Czechia, Slovakia, Poland and Hungary, NORDIS in Finland, Sweden, Norway and Denmark, and EDMO BELUX in Belgium and Luxembourg. This network is joined by European fact checker organizations with whom they have established an agreement to promote the detection and neutralization of disinformation related to the war in Ukraine.

The check It is a business that moves a lot of capital considering that most of it is private and without losing sight of the fact that misinformation is very inexpensive and its elimination means big investments.

Another example of an international agreement is found in Brazil. Sergio Lüdtke, project editor check, explains to us in Oslo that this coalition brings together about 40 media outlets, members of the Brazilian Association of Investigative Journalism (Abraji, for its acronym in Portuguese).

It is striking that they all follow the same method: the verifications are put in a document shared between several journalists in the coalition who begin to work on the verification. After discussing it, functions are assigned such as going to look for the information, who will respond, the original sources are sought, those responsible for publishing the rumor for the first time. When it has been completed, a cross-verification is proposed since, according to Sergio Lüdtke, before being published, the content must be verified by three other media that did not participate in the original verification.

Behind this initiative is Google y Metaverso. They subsidize the media so that they can have three of their workers specialize in the task of verifying. At the end of the day, verification is a business that moves a lot of capital taking into account that most of it is private and without losing sight of the fact that disinformation is very inexpensive and its elimination requires large investments.

---

Against the clock

Precisely for economic reasons, the time it takes to verify matters. The general director of Checked (project of the La Voz Pública foundation, a non-profit digital media that verifies public discourse and works for the verification and opening of data), Laura Sommers: “We fact checkers around the world are very clear that time counts, that "The sooner we react to the lie, the more chances we have." Checked It started 12 years ago as the first initiative in the Global South.

In this sense, the generation of alliances from the project is very necessary. IVERES, since the more agents, media and entities we are united in this fight, the more effective the enormous work they carry out will be. hundreds of verifiers from all over the world. As far as the critical analysis of mature environments is concerned, IVERES has been of great interest to all of us who have asked because they would benefit from having a useful tool that combines current programs for monitoring and analyzing information on social networks, image verification, text , detection of deep fake audio and video, with the final objective of helping journalists, verifiers and official institutions, if possible, to put end to the rapid expansion of possible hoaxes on networks And power debunk malicious disinformation campaigns.

Montserrat Rigall

Journalist from RTVE.

Other articles on RTVE