The disinformation sleuths: a key role for scientists in impending elections

Next year will bring a series of high-profile elections around the globe, including in India, Taiwan, the United States and, in all likelihood, the United Kingdom, as well as for the European Parliament. Social media will play a huge part in bringing information to the hundreds of millions of people casting their votes — and researchers who study elections are worried.

Access to social-media data is essential to those who research political campaigns and their outcomes. However, unlike in previous years, scientists will not have free access to data from X, previously known as Twitter. Many still consider X to be among the world’s most influential social-media platforms for political discussion, but the company has discontinued its policy of giving researchers special access to its data. Disinformation campaigns — some armed with AI-generated deepfakes — are likely to be rampant in the coming months, says Ulrike Klinger, who studies political communication at the European University Viadrina in Frankfurt (Oder), Germany. “And we cannot monitor them because we don’t have access to data.”

Researchers scramble as Twitter plans to end free data access

Until its change of policy, X was an outlier in its open approach to providing data for research. Because researchers’ access to data from technology platforms is controlled by the companies themselves, firms can cherry-pick which studies they allow to go forwards, potentially creating a skewed image of their performance.

Tech companies are starting to report on how they are tackling online harms, as many did last week in submissions to the European Union’s Transparency Centre. But good science demands studies from individuals and teams unaffiliated with the platforms. Such studies would make it possible to authenticate the claims made in the companies’ reporting, or to determine how common misinformation is, which communities are being targeted, and how effective — or harmful — that misinformation is. Beyond the immediate concerns about elections, reliable data are also needed to address long-standing concerns about online platforms, including their impact on mental health, and the prevalence of harassment, privacy violations and hate speech associated with gender, ethnicity, sexuality and other characteristics.

The EU at least is making the right moves. Its Digital Services Act was agreed in 2022. The bulk of its provisions are due to apply from early next year, and should ensure that very large online platforms — those with more than 45 million users — open up relevant data to vetted researchers deemed to be independent of commercial interests. That should crack open a treasure chest of data from social media, search engines such as Google and e-commerce platforms such as Amazon.

How to stop AI deepfakes from sinking society — and science

But such work will require the EU regulation to be implemented fairly, which is not a given. A competent authority in each country, called a Digital Services Coordinator, will mediate researchers’ requests for data. This marks a big change, because it means researchers will not be beholden to the whims of companies. But a company can still refuse to provide data or ask for amendments to the request, for example, if it thinks that its data might not be secure in researchers’ hands, and that confidential information, such as trade secrets, could be divulged. And each EU member state is free to interpret security, confidentiality and trade secrets according to its own laws. If a definition is too broad, it could lead to many, if not most, requests being denied.

This is why researchers need to step up and work with policymakers to define the procedures for determining which data risk divulging trade secrets, how best to ensure equitable access to data and how data quality should be assessed, so that disputes can be quickly resolved. If this doesn’t happen, requests for data could be tied up, possibly in litigation, for so long that the information ceases to be useful.

The more details that can be laid out at the start, the less opportunity there will be for companies to exploit a lack of clarity to delay or contest requests for data. Researchers must speak up now to ensure that their needs — and the goals of the Digital Services Act to provide independent assessments of platforms’ impact on society — are met.

If democratic societies are to thrive, it is essential that independent researchers have the legal right to access online data and study them without interference from the companies to which the data belong. In an ominous sign of what might lie ahead, X’s owner, Elon Musk, last week confirmed rumours that he had disbanded the platform’s Election Integrity Team. This was set up by the previous owners in an effort to stop the platform being exploited to cause harm during elections. Musk’s move makes it all the more important that other platforms work constructively with researchers to ensure that both the letter and the spirit of the EU Digital Services Act are implemented.