26 September 2024

DisinfoCon

DRI’s DisinfoCon is a forum for civil society, policymakers, journalists and AI practitioners to meet, exchange and align on a values-based approach to the AI revolution affecting our democracies. This year’s DisinfoCon will focus on the landscape of disinformation in this pivotal election year and solutions to protect information integrity.

Highlights

Our data-driven analyses provide updates on the ever-evolving disinformation landscape.

ChatGPT vs. Bard: Unveiling the Battle against Disinformation and Creative Output

ChatGPT vs. Bard: Unveiling the Battle against Disinformation and Creative Output

Through testing and comparison of the generative AI chatbots ChatGPT and Bard in the following dimensions: (i) prevention, (ii) circumvention and (iii) creativity, we observe insights into the chatbots’ potential vulnerabilities, their responses to misleading information, and the extent to which they can be manipulated to propagate disinformation.

An Ambivalent Alliance: How Authoritarian Regimes Use – and Fear –Generative AI

An Ambivalent Alliance: How Authoritarian Regimes Use – and Fear –Generative AI

Generative AI models are now being developed in authoritarian states and have the potential to make regulation in democratic states ineffective. The user-friendly nature of these models creates a proliferation risk – they may spread their intentional biases beyond the borders of their countries of origin.

From Prompt to Problematic: How ChatGPT Reproduces Country-Specific Misinformation

From Prompt to Problematic: How ChatGPT Reproduces Country-Specific Misinformation

ChatGPT can easily be used to generate propagandistic and harmful narratives in different languages around the world. To investigate this, we tested three country-specific narratives in three different languages (Portuguese, English, and German), attempting to circumvent the chatbot’s safety restrictions.

Going Beyond the Radar: Emerging Threats, Emerging Solutions

Going Beyond the Radar: Emerging Threats, Emerging Solutions

Understanding how disinformation campaigns unfold in different country contexts and what are counterstrategies that (non-)governmental stakeholders develop to tackle disinformation on a local level is essential to fight information manipulation holistically.

Worth More than 1,000 words: The Disinformation Potential of Text-to-Video Generators

Worth More than 1,000 words: The Disinformation Potential of Text-to-Video Generators

Innovations in AI-powered image creation have reached a fever pitch this year. New technologies that allow users to create realistic images from simple text prompts, such as OpenAI’s DALL-E 2, Google Brain’s Imagen, and Stability AI’s Stable Diffusion, brought with them distinct new potential for disinformation.

Is AI Undermining Trust Online? ChatGPT, Large Language Models, and Disinformation 

Is AI Undermining Trust Online? ChatGPT, Large Language Models, and Disinformation 

ChatGPT has caused quite a stir on social media and has been used more than 1 million times in just five days. However, what happens when ChatGPT mixes false and correct information or fabricates disinformation? How will users distinguish between a text produced by a human and an advanced language model?

WhatsApps’s Paradox Reality: Disinformation Tactics during the 2022 Brazilian Elections

WhatsApps’s Paradox Reality: Disinformation Tactics during the 2022 Brazilian Elections

Since 2018, Brazil has seen the development of a true disinformation ecosystem, characterised by powerful actors, mass messaging, and the production and dissemination of falsified content. This disinformation ecosystem has been responsible for an avalanche of false information that confuses the population and impacts institutions.

Pay to Pray: The Privacy Pitfalls of Faith-based Mobile Apps

Pay to Pray: The Privacy Pitfalls of Faith-based Mobile Apps

Investigations into several faith-based apps reveal that many harvest sensitive information and sell it to untransparent third-party vendors. Such practices have the potential to provide malicious actors with the information they need to craft targeted disinformation campaigns.

Stable Diffusion, Open Access Image Generation and Disinformation

Stable Diffusion, Open Access Image Generation and Disinformation

Stable Diffusion is one of the most powerful AI text-to-image generators ever developed. It is freely available and comes with few restrictions. That raises ethical questions – especially regarding racist or sexually explicit content.

What a Pixel Can Tell: Text-to-Image Generation and its Disinformation Potential

What a Pixel Can Tell: Text-to-Image Generation and its Disinformation Potential

Text-to-image-generation is an impressive technology with fascinating potential and legitimate uses. But it could have daunting effects on our democratic public discourse.

On the Radar: Mapping the Tools, Tactics and Narratives of Tomorrow’s Disinformation Environment

On the Radar: Mapping the Tools, Tactics and Narratives of Tomorrow’s Disinformation Environment

The hazy world of online disinformation is about to get even hazier. In this publication, DRI analyses the cutting-edge tools and strategies shaping the future of disinformation.

The 2021 German federal election in retrospect: A meta analysis of the online election campaign

The 2021 German federal election in retrospect: A meta analysis of the online election campaign

Germany’s September 2021 federal election took place in a highly online environment during the pandemic, increasing risks of digital violence against politicians, negative campaigning and attempts at coordinated online disinformation.