This database systematizes public information on influence actions, bills, and litigation involving major technology companies in different countries and jurisdictions. Its aim is to provide an overview of the strategies these corporations deploy to shape the political agenda and regulatory frameworks that affect their activities.
It is not an exhaustive record of all influence activities in the countries covered by the investigation, since the content depends on the quality and level of detail of the publicly available information in each country.
This database may be of interest to journalists, researchers, academics, policymakers, civil society actors, and individuals seeking to understand how tech companies use these tactics to influence public decisions that affect them, and what issues concern them the most.
Influence strategies are not always easy to detect, partly because not all countries require them to be disclosed. And even in those that do, the information available to the public often requires additional context to be properly interpreted. For this reason, the Latin American Center for Investigative Journalism (CLIP), together with Agência Pública (Brazil) and 15 other media outlets, collected and standardized influence actions across our countries.
Each reporting team searched for and contributed, to the extent that information was available, three types of data from its country:
- Actions taken by tech companies to influence public decisions—particularly meetings and various interactions between their employees, lawyers, or associations and public officials from the executive and legislative branches.
- Bills seeking to regulate these companies or affecting their interests.
- Litigation or administrative actions before courts or regulatory bodies in which these companies appear as plaintiffs or defendants.
To bring together thousands of records of different types in a single standardized repository, we categorized them by subject using artificial intelligence; manually verified information related to the identities of people involved in influence actions; and carried out additional verification with project partners.
The result is five datasets documenting:
- Influence actions (2,977) targeting officials in Argentina, Australia, Brazil, Canada, Chile, Colombia, Costa Rica, Ecuador, El Salvador, the United States, Guatemala, Kenia, Mexico, Paraguay, Peru, the Dominican Republic, Uruguay, and a supranational entity (the European Union). These were carried out by the Big Five—Google, Apple, Meta (owner of Facebook), Amazon, and Microsoft—their allied companies, other technology firms with major impact on people’s daily lives (such as Mercado Libre and TikTok), and lobbying groups and associations. We also included key firms associated with data centers. This dataset is called “influence_actions.”
- Bills (801) from Argentina, Australia, Brazil, Canada, Chile, Colombia, Ecuador, El Salvador, Indonesia, Mexico, Paraguay, and the European Union. This dataset is called “bills.”
- Court cases (315) promoted by or against tech companies in Australia, Brazil, Canada, Chile, and Colombia. This dataset is called “court_cases.”
- Public officers (2,508) who signed bills listed in dataset #2 or who participated in influence actions listed in dataset #3. This dataset is called “public_officers.”
- People who took part in influential actions. This list includes 1.516 people who took part in at least one of the meetings listed in the dataset #1. This dataset is called “tech_reps.” Of them:
- 69 are representatives from civil society organizations.
- 32 are representatives from other kinds of organizations. (for example, guilds or trade associations of sectors other than technology, international or multilateral organizations, or academics).
- 1.414 are representatives from the technology industry. They can be direct representatives from trade companies, representatives from other companies that support tech companies, and representatives from law firms or PR agencies who act on behalf of technology companies.
The databases are written in English.
Data Sources and Inclusion Criteria
The sources from which the data were obtained are public. Some come from official information, such as government transparency portals or institutional entry records; others from publicly accessible information, such as press releases and official statements. More than 180 sources were used. The complete list can be found at the end of this methodological note.
Influence Actions
For the dataset “influence_actions,” most data come from official government platforms, public online registries of lobbying and special-interest management activities in countries where such registries exist—such as the Infolobby portal in Chile and the Agenda Transparente platform in Brazil. It should be noted that the actions recorded in these portals may omit meetings and encounters held outside working hours or outside institutional premises (for example, in a restaurant or a bar). This is the case with Infolobby, since in Chile it is not mandatory to record meetings that take place in informal settings.
In countries where there is no obligation to record private-sector lobbying actions before public institutions, this journalistic alliance submitted multiple public information requests to various state entities seeking details about tech company operators; consulted information published on social media and in news outlets; and obtained other public records such as entries into government buildings or conflict-of-interest declarations, among others. Evidence was also drawn from historical press archives, civil society organizations, think tanks, and corporate platforms.
The information about these meetings and the names of participants was manually verified by journalists in this alliance at different stages of building the database. In the url column of the dataset, links can be found to the primary source of information for each specific influence action.
This alliance has defined “influence actions” as meetings (and meeting requests) between representatives of technology companies and public officials; hearings; protocol greetings; courtesy visits; the delivery of gifts; event invitations; revolving-door cases; trainings offered by tech companies to public institutions to showcase their products; company statements on bills; and events where, based on their description, it is possible to deduce there was interaction between public officials and employees or representatives of tech companies.
We also included records of actions whose specific nature could not be determined—for example, entries without descriptions or events with scant information. These were incorporated because they involve at least one representative of a tech company and one public official, or because a tech company is mentioned in the description of the encounter.
In Brazil, 904 lobbyist entries to the Chamber of Deputies were recorded between 2021 and 2025. However, these data are not included in all visualizations because the team was unable to independently verify the purpose of each visit. In the repository, these actions appear tagged with “no clear topic.”
The database includes individuals whose participation in these encounters we were able to document: public officials, with their names and positions; and tech industry representatives, with their names and the companies they represent. This information was corroborated with different sources. The names of other tech executives—though they appear in different documents—were excluded whenever we could not verify their participation in influence actions.
It should be stressed that taking part in one of these meetings, or appearing in this or other databases that document influence actions in the public sector, is not in itself an illegal or illegitimate act. However, we consider the publication of this information to be in the public interest, since it relates to public decision-making processes and makes transparent an exercise of influence in public affairs.
Similarly, it should be noted that the country associated with each influence action corresponds to that of the public officials involved in the action.
Finally, it is important to note that in some cases, errors were detected in names obtained from the Infolobby platform. During the verification process, reasonable efforts were made to identify and correct these errors, but this alliance cannot guarantee that all of them have been resolved. We also found that the same person may be registered multiple times on the platform with variations in their name. For these reasons, some names included in the database may contain inaccuracies derived directly from this or other official sources and original documents.
Bills
The bills compiled in the “bills” dataset come from official sources of parliaments and legislative platforms, including the digital repositories of national senates and congresses, as well as portals tracking bills, provincial legal databases, and jurisprudence and regulatory platforms. They were selected because they deal with issues of technology regulation, artificial intelligence, digital security, or social media.
In the case of legislative initiatives, the alliance decided to keep visible the names of the legislators and officials who sign them, since this is public information available in the official repositories of Congresses and Parliaments. We make these names public on principles of transparency, accountability, and public interest, as they are considered essential for analyzing the legislative process and do not compromise the personal privacy of the actors involved.
Court Cases
Finally, for the “court_cases” dataset, the information used comes from official and open-access repositories, including rulings, resolutions, and procedural filings from courts in Australia (AustLII), Brazil, Canada (Supreme Court), Colombia (Council of State), and Chile (Environmental Court and Court For Free Competition). In Canada, for one specific case, the information was obtained from an announcement by the social network TikTok and verified on the Federal Court of Vancouver’s website. These cases were selected because they mention a technology company as plaintiff or defendant, or because the company is referenced in the description of the proceeding.
Use of Artificial Intelligence to Categorize by Topic
Once the information was gathered, the alliance sought to identify common themes in lobbying actions. For this exercise, artificial intelligence (AI) was used.The AI was provided with the subject, tags, and url (or link) columns from the three datasets, in order to analyze the information and generate an initial list of common topics. This list served as the basis for another list, compiled manually by journalists. The result was a classification into 21 topics.
| Tema (em inglês) | Descrição |
| AI | Regulation, adoption, and use of AI. Everything explicitly related to AI. |
| Children’s Rights | Issues specifically related to children: protection, risks, access to platforms, etc. |
| Data Centers | Issues related to the construction and operation of data centers, as well as their social and environmental impact. |
| News Media | Bargaining codes, regulations, and negotiations regarding media content on platforms. |
| Personal Data & Privacy | Issues related to privacy and personal data regulation. |
| Taxes | Tax policy, tax exemptions. |
| Elections | Electoral advertising, disinformation in elections, rules on electoral and political content. |
| Scams & Fraud | Scams, fraud, petty theft committed through digital platforms. |
| Content Regulation | Rules on content in digital platforms: content regulation, distribution, monetization, etc. |
| Content Moderation | Rules on moderation, i.e., the internal application of content regulation: the moderation industry, transparency in moderation, work of moderators. |
| Antitrust / Competition | IInteractions related to antitrust or competition regulations, cases, or decisions. |
| Public Procurement | Procurement, bids, or commercial interactions between tech companies and the public sector. |
| E-Commerce | Issues related to e-commerce: logistics, payments, digital marketing, etc. |
| Energy | Energy issues: construction of energy infrastructure, matters related to energy availability, etc. |
| Environment & Sustainability | Environmental and climate issues (water, climate, carbon emissions, “clean” technologies, etc.). |
| International Affairs | Geopolitics, international affairs, diplomacy. |
| Health | Health services, regulation of health technologies. |
| Labor & Employment | Labor issues: platform work, matters related to labor contracts, labor regulation. |
| Security & Defense | Technology for defense and the military, technology for public security, cybersecurity, public security issues except scams and fraud (e.g., drug trafficking, organized crime). |
| Education | Educational issues: technology education, adoption of technology in schools/universities. |
| Social Development | General social policy issues (e.g., poverty, food, agriculture). |
| Tech infrastructure | All other issues related to technological infrastructure, except data centers: cables, network construction, satellites, fiber optics, spectrum, Wi-Fi, digital transformation, etc. |
| Government & Policy | E-government, general policy proposals. |
Based on this framework, scripts were programmed for the AI to assign between one and three topics to each lobbying action, bill, or court case. After the first assignment, all records were manually reviewed; some thematic categories were refined and adjusted, and the AI was asked to perform a second reclassification. In the case of Brazil, which had the largest number of records, a second round of verification by journalists was carried out to produce the final version.
Even then, it was found that some records still could not be clearly classified, due to the limited information available in public sources (for example, in Brazil there were entries with very general descriptions such as “meeting” with a tech company). These cases were grouped under the category “no clear topic.” In total, 1,245 actions without a defined topic were identified.
For this reason, these cases were excluded from visualizations showing data by topic, although they were retained in the database to reflect the recorded activity—except in those cases where it was explicitly stated that the event never took place.
It is also important to note that a single influence action may be classified under multiple topics. This is why thematic charts do not display the exact number of actions, but rather how many times topics are repeated.
Considerations for Using the Database
The database allows users to identify common strategies and topics across different jurisdictions. It can be useful for detecting patterns in how tech companies defend their interests, and for observing the issues that are most sensitive to them and to societies in relation to platform use in different countries.
That said, we believe there are three aspects that anyone consulting this database should keep in mind when conducting their analyses or citing it as a source:
- The database is not exhaustive and does not provide the full scope of interactions between big tech companies and their allies with governments in these countries and jurisdictions. It depends on the legislation of each country—whether it requires the registration of influence actions—the degree of transparency required by law, and the quality of its records.
- Legislation and levels of transparency vary across countries. This means that the volume and detail of information are not directly comparable between them.
- The inclusion of proper names of legislators, officials, and company representatives serves transparency purposes. These names were obtained from public and open-access records and sources. The fact that they appear promoting their interests before a public official, or interacting with representatives of the tech industry, is part of a legitimate function in a democracy and does not imply any illegal or illegitimate conduct. Also, the people who are featured in this list as representatives or employees from civil society organizations, or from other sectors, should not be considered representatives from tech companies.
- This journalistic team performed a manual verification to avoid inaccuracies arising from errors in the source data. However, some errors may remain. We would appreciate it if you could report them to jpenarredonda@elclip.org.
Links to Generic Sources
The exact sources for each record can be consulted in the downloadable database, in the columns titled “url” or “link to document” of each of the three datasets.
Dataset “Influence Actions”
Brasil
- https://acervo.estadao.com.br/
- https://agendas.fiquemsabendo.com.br/
- https://adrianaventura.com/
- https://brasscom.org.br/
- https://blog.youtube/intl/pt-br/inside-youtube
- https://www.boletimdaliberdade.com.br/
- https://conselhodigital.org.br/
- https://www.camara.leg.br/fale-conosco
- https://www.camara.leg.br/proposicoesWeb/fichadetramitacao
- https://www2.camara.leg.br/a-camara/estruturaadm/mesa/segunda-secretaria/servicos/merito-legislativo/Agraciados2022Parapublicarfinal.pdf
- https://capitaldigital.com.br/
- https://www.congressoemfoco.com.br/
- https://eagendas.cgu.gov.br/
- https://elections.registro.br/
- https://about.fb.com/br/news
- https://www1.folha.uol.com.br/
- https://gazetadocerrado.com.br/
- https://www.instagram.com/eduardogomestocantins/
- https://www.instagram.com/fppelobrasilcompetitivo/?hl=es
- https://www.instagram.com/luisa_canziani/?hl=es
- https://legis.senado.leg.br
- https://www.linkedin.com/in/roberta-rios/
https://www.linkedin.com/in/marcelolacerda1973/ - https://www.linkedin.com/in/marcio-gon%C3%A7alves-9394922/?locale=en_US
- https://www.linkedin.com/in/francoismartins/
- https://www.linkedin.com/in/ricardo-leite-ribeiro-b9ab22113/
- https://www.linkedin.com/in/monica-steffen-guise/?originalSubdomain=br
- https://www.linkedin.com/in/andre-atadeu/
- https://www.linkedin.com/in/rafael-alloni/
- https://www.linkedin.com/in/kalianakalache/
- https://www.linkedin.com/in/joao-barroso/
- https://www.linkedin.com/in/marcos-tourinho-31690517/?originalSubdomain=br
- https://www.linkedin.com/in/joao-sabino-392a6aa9/
- https://www.linkedin.com/in/michael-beckerman-9b750a58/
- https://www.linkedin.com/in/mario-vilhena/
- https://www.linkedin.com/in/flavia-annenberg/
- https://www.linkedin.com/in/gabriela-silveira-50373740/
- https://www.linkedin.com/in/fernanda-laranja-2b6a1992/
- https://www.linkedin.com/in/pedroless/
- https://www.linkedin.com/in/yana-dumaresq-sobral-a-37a19429/
- https://www.linkedin.com/in/fabianobarreto/
- https://www.linkedin.com/in/analucordeiro/
- https://www.linkedin.com/in/gabrielabarbosa/
- https://www.linkedin.com/in/guilherme-horn/
- https://www.linkedin.com/in/marcocarneiro-corporaterelations/?locale=en_US
- https://www.linkedin.com/in/murillolaranjeira/
- https://www.linkedin.com/in/murillolaranjeira/
- https://www.linkedin.com/in/luiz-fernando-marrey-moncau/
- https://www.linkedin.com/in/beatriz-sannuti-de-carvalho-6a782222/?originalSubdomain=br
- https://www.linkedin.com/in/debora-gershon-9b3bb6ba/
- https://www.linkedin.com/in/julianambueno/
- https://www.linkedin.com/in/roberta-rios/
- https://www.linkedin.com/in/analuizavaladares/
- https://www.linkedin.com/in/cibeleperillo/
- https://www.linkedin.com/in/felipedaud/
- https://www.linkedin.com/in/mariana-polidorio/
- https://www.linkedin.com/in/paulafarani/
- https://www.linkedin.com/in/michael-mohallem/?locale=en_US
- https://www.linkedin.com/in/paulafarani?originalSubdomain=br
- https://www.linkedin.com/in/irinacezar/
- https://www.linkedin.com/in/fernando-paes-714263180/
- https://www.linkedin.com/in/sophia-martini-vial-8370278a/
- https://www.linkedin.com/in/bruna-lopes-ab6135146/
- https://www.linkedin.com/in/gabriela-sanches-544717244/
- https://www.linkedin.com/in/isabelaaguiar1/?originalSubdomain=br
- https://www.linkedin.com/in/mariodelacruzsarabia/
- https://www.linkedin.com/in/grecalde/?originalSubdomain=br
- https://www.linkedin.com/in/josiane-trevisan/
- https://www.linkedin.com/in/wanderleymariz/
- https://www.linkedin.com/in/nicolasrobinsonandrade/
- https://www.linkedin.com/in/marcelatullii/?originalSubdomain=br
- https://www.linkedin.com/in/ivo-correa-tech/
- https://www.linkedin.com/in/husanidurans/
- https://www.linkedin.com/in/pablobello/?originalSubdomain=br
- https://www.linkedin.com/in/maria-eduarda-cintra-96069490/
- https://www.linkedin.com/in/josiaradiniz/
- https://www.metropoles.com/
- https://noticias.uol.com.br/
- https://nucleo.jor.br/
- https://www.poder360.com.br/
- https://reglab.com.br/
- https://static.portaldaindustria.com.br/
- https://www.sympla.com.br/
- https://web.archive.org/
- https://www.youtube.com/@MinFazenda
Canadá
- https://blog.google/
- https://canada2020.ca/
- https://lobbycanada.gc.ca/
- https://nationalpost.com/
- https://sencanada.ca/
- https://theijf.org/
- https://thelogic.co/
- https://www.linkedin.com/in/rachel-curran-a99258109/?originalSubdomain=ca
- https://www.pm.gc.ca/
- https://www.ourcommons.ca/
- https://www.theglobeandmail.com/
- https://www.theguardian.com/
- https://www.youtube.com/@JJsSeconds
Chile
Colômbia
- https://expoicolombia.com/agenda-academica/
- https://impactotic.co/
https://leyes.senado.gov.co/ - https://www.alianzaincongreso.com/
- https://www.camara.gov.co/
- https://www.elheraldo.co/
- https://www.linkedin.com/company/centromexicodigital/posts/?feedView=all
- https://www.linkedin.com/company/asociaci-n-latinoamericana-de-internet/posts/?feedView=all
- https://www.linkedin.com/in/germ%C3%A1n-l%C3%B3pez-ardila-52084845/
- https://www.linkedin.com/in/pablo-nieto-d-719692b3/
- https://www.linkedin.com/in/david-luna-b524a0174/
- https://www.uexternado.edu.co/
- https://www.youtube.com/@CanalCongresoColombia
- https://www.youtube.com/@comisionprimera
- https://www.youtube.com/@comisiontercera4970
- https://www.youtube.com/@comisionsextadesenado5710
- https://www.youtube.com/@comisionsexta-camaraderepr8761
https://www.youtube.com/@mesacolombianadegobernanza4161 - https://www.youtube.com/@superintendenciaSIC
Ecuador
- https://transparencia.dpe.gob.ec/entidades/1176#
- https://colabora.asambleanacional.gob.ec/
- https://www.facebook.com/watch/ComisionSoberaniaAN/
- https://www.facebook.com/Niuboxlegal
- https://www.instagram.com/citecec/?hl=es
- https://www.linkedin.com/company/centromexicodigital/posts/?feedView=all
- https://educacion.gob.ec/
- https://jprf.gob.ec/
- https://vlex.ec/
- https://citec.com.ec/
- https://www.asambleanacional.gob.ec/
Paraguai
República Dominicana
União Europeia
- https://ec.europa.eu
- https://x.com/KenInvest
- https://newsroom.gy/
- https://en.antaranews.com/
- https://en.tempo.co/
- https://www.stabroeknews.com/
- https://health.gov.gy/
- https://archive.is/
- https://web.archive.org/
- https://www.thejakartapost.com/
- https://observerid.com/
- https://x.com/tanuijohn
- https://teknologi.bisnis.com/
Dataset “Bills”
Canadá
El Salvador
Indonésia
Paraguai
União Europeia
Big Tech’s Invisible Hand is a cross-border, collaborative journalistic investigation led by Brazilian news organization Agência Pública and the Centro Latinoamericano de Investigación Periodística (CLIP), together with Crikey (Australia), Cuestión Pública (Colombia), Daily Maverick (South Africa), El Diario AR (Argentina), El Surti (Paraguay), Factum (El Salvador), ICL (Brazil), Investigative Journalism Foundation – IJF (Canada), LaBot (Chile), LightHouse Reports (International), N+Focus (Mexico), Núcleo (Brazil), Primicias (Ecuador), Tech Policy Press (USA), and Tempo (Indonesia). Reporters Without Borders and the legal team El Veinte supported the project, and La Fábrica Memética designed the visual identity.
PayPal 
