Government & Policy

EU publishes election security guidance for social media giants and others in scope of DSA

Comment

The human hand drops the ballot into the box.
Image Credits: Boris Zhitkov / Getty Images

The European Union published draft election security guidelines Tuesday aimed at the around two dozen (larger) platforms with more than 45 million+ regional monthly active users who are regulated under the Digital Services Act (DSA) and — consequently — have a legal duty to mitigate systemic risks such as political deepfakes while safeguarding fundamental rights like freedom of expression and privacy.

In-scope platforms include the likes of Facebook, Google Search, Instagram, LinkedIn, TikTok, YouTube and X.

The Commission has named elections as one of a handful of priority areas for its enforcement of the DSA on very large online platforms (VLOPs) and very large online search engines (VLOSEs). This subset of DSA-regulated companies are required to identify and mitigate systemic risks, such as information manipulation targeting democratic processes in the region, in addition to complying with the full online governance regime.

Per the EU’s election security guidance, the bloc expects regulated tech giants to up their game on protecting democratic votes and deploy capable content moderation resources in the multiple official languages spoken across the bloc — ensuring they have enough staff on hand to respond effectively to risks arising from the flow of information on their platforms and act on reports by third-party fact-checkers — with the risk of big fines for dropping the ball.

This will require platforms to pull off a precision balancing act on political content moderation — not lagging on their ability to distinguish between, for example, political satire, which should remain online as protected free speech, and malicious political disinformation, whose creators could be hoping to influence voters and skew elections.

In the latter case, the content falls under the DSA categorization of systemic risk that platforms are expected to swiftly spot and mitigate. The EU standard here requires that they put in place “reasonable, proportionate, and effective” mitigation measures for risks related to electoral processes, as well as respecting other relevant provisions of the wide-ranging content moderation and governance regulation.

The Commission has been working on the election guidelines at pace, launching a consultation on a draft version just last month. The sense of urgency in Brussels flows from upcoming European Parliament elections in June. Officials have said they will stress-test platforms’ preparedness next month. So the EU doesn’t appear ready to leave platforms’ compliance to chance, even with a hard law in place that means tech giants are risking big fines if they fail to meet Commission expectations this time around.

User controls for algorithmic feeds

Key among the EU’s election guidance aimed at mainstream social media firms and other major platforms are that they should give their users a meaningful choice over algorithmic and AI-powered recommender systems — so they are able to exert some control over the kind of content they see.

“Recommender systems can play a significant role in shaping the information landscape and public opinion,” the guidance notes. “To mitigate the risk that such systems may pose in relation to electoral processes, [platform] providers … should consider: (i.) Ensuring that recommender systems are designed and adjusted in a way that gives users meaningful choices and controls over their feeds, with due regard to media diversity and pluralism.”

Platforms’ recommender systems should also have measures to downrank disinformation targeted at elections, based on what the guidance couches as “clear and transparent methods,” such as deceptive content that’s been fact-checked as false and/or posts coming from accounts repeatedly found to spread disinformation.

Platforms must also deploy mitigations to avoid the risk of their recommender systems spreading generative AI-based disinformation (aka political deepfakes). They should also be proactively assessing their recommender engines for risks related to electoral processes and rolling out updates to shrink risks. The EU also recommends transparency around the design and functioning of AI-driven feeds and urges platforms to engage in adversarial testing, red-teaming, etc., to amp up their ability to spot and quash risks.

On GenAI the EU’s advice also urges watermarking of synthetic media — while noting the limits of technical feasibility here.

EU dials up scrutiny of major platforms over GenAI risks ahead of elections

Recommended mitigating measures and best practices for larger platforms in the 25 pages of draft guidance published today also lay out an expectation that platforms will dial up internal resourcing to focus on specific election threats, such as around upcoming election events, and putting in place processes for sharing relevant information and risk analysis.

Resourcing should have local expertise

The guidance emphasizes the need for analysis on “local context-specific risks,” in addition to member state specific/national and regional information gathering to feed the work of entities responsible for the design and calibration of risk mitigation measures. And for “adequate content moderation resources,” with local language capacity and knowledge of the national and/or regional contexts and specificities — a long-running gripe of the EU when it comes to platforms’ efforts to shrink disinformation risks.

Another recommendation is for them to reinforce internal processes and resources around each election event by setting up “a dedicated, clearly identifiable internal team” ahead of the electoral period — with resourcing proportionate to the risks identified for the election in question.

The EU guidance also explicitly recommends hiring staffers with local expertise, including language knowledge. Platforms have often sought to repurpose a centralized resource — without always seeking out dedicated local expertise.

“The team should cover all relevant expertise including in areas such as content moderation, fact-checking, threat disruption, hybrid threats, cybersecurity, disinformation and FIMI [foreign information manipulation and interference], fundamental rights and public participation and cooperate with relevant external experts, for example with the European Digital Media Observatory (EDMO) hubs and independent factchecking organisations,” the EU also writes.

The guidance allows for platforms to potentially ramp up resourcing around particular election events and de-mobilize teams once a vote is over.

It notes that the periods when extra risk mitigation measures may be needed are likely to vary, depending on the level of risks and any specific EU member state rules around elections (which can vary). But the Commission recommends that platforms have mitigations deployed and up and running at least one to six months before an electoral period, and continue at least one month after the elections.

Unsurprisingly, the greatest intensity for mitigations is expected in the period prior to the date of elections, to address risks like disinformation targeting voting procedures.

Hate speech in the frame

The EU is generally advising platforms to draw on other existing guidelines, including the Code of Practice on Disinformation and Code of Conduct on Countering Hate Speech, to identify best practices for mitigation measures. But it stipulates they must ensure users are provided with access to official information on electoral processes, such as banners, links and pop-ups designed to steer users to authoritative info sources for elections.

“When mitigating systemic risks for electoral integrity, the Commission recommends that due regard is also given to the impact of measures to tackle illegal content such as public incitement to violence and hatred to the extent that such illegal content may inhibit or silence voices in the democratic debate, in particular those representing vulnerable groups or minorities,” the Commission writes.

“For example, forms of racism, or gendered disinformation and gender-based violence online including in the context of violent extremist or terrorist ideology or FIMI targeting the LGBTIQ+ community can undermine open, democratic dialogue and debate, and further increase social division and polarization. In this respect, the Code of conduct on countering illegal hate speech online can be used as inspiration when considering appropriate action.”

It also recommends they run media literacy campaigns and deploy measures aimed at providing users with more contextual info — such as fact-checking labels; prompts and nudges; clear indications of official accounts; clear and non-deceptive labeling of accounts run by member states, third countries and entities controlled or financed by third countries; tools and info to help users assess the trustworthiness of information sources; tools to assess provenance; and establish processes to counter misuse of any of these procedures and tools — which reads like a list of stuff Elon Musk has dismantled since taking over Twitter (now X).

Notably, Musk has also been accused of letting hate speech flourish on the platform on his watch. And at the time of writing, X remains under investigation by the EU for a range of suspected DSA breaches, including in relation to content moderation requirements.

Transparency to amp up accountability

On political advertising, the guidance points platforms to incoming transparency rules in this area — advising they prepare for the legally binding regulation by taking steps to align themselves with the requirements now. (For example, by clearly labeling political ads, providing information on the sponsor behind these paid political messages, maintaining a public repository of political ads, and having systems in place to verify the identity of political advertisers.)

Elsewhere, the guidance also sets out how to deal with election risks related to influencers.

Platforms should also have systems in place enabling them to demonetize disinformation, per the guidance, and are urged to provide “stable and reliable” data access to third parties undertaking scrutiny and research of election risks. Data access for studying election risks should also be provided for free, the advice stipulates.

More generally the guidance encourages platforms to cooperate with oversight bodies, civil society experts and each other when it comes to sharing information about election security risks — urging them to establish comms channels for tips and risk reporting during elections.

For handling high-risk incidents, the advice recommends platforms establish an internal incident response mechanism that involves senior leadership and maps other relevant stakeholders within the organization to drive accountability around their election event responses and avoid the risk of buck passing.

Post-election, the EU suggests platforms conduct and publish a review of how they fared, factoring in third-party assessments (i.e., rather than just seeking to mark their own homework, as they have historically preferred, trying to put a PR gloss atop ongoing platform manipulated risks).

The election security guidelines aren’t mandatory, as such, but if platforms opt for another approach than what’s being recommended for tackling threats in this area, they have to be able to demonstrate their alternative approach meets the bloc’s standard, per the Commission.

If they fail to do that, they’re risking being found in breach of the DSA, which allows for penalties of up to 6% of global annual turnover for confirmed violations. So there’s an incentive for platforms to get with the bloc’s program on ramping up resources to address political disinformation and other info risks to elections as a way to shrink their regulatory risk. But they will still need to execute on the advice.

Further specific recommendations for the upcoming European Parliament elections, which will run June 6–9, are also set out in the EU guidance.

On a technical note, the election security guidelines remain in draft at this stage. But the Commission said formal adoption is expected in April once all language versions of the guidance are available.

EU’s draft election security guidelines for tech giants take aim at political deepfakes

More TechCrunch

For over six decades, the nonprofit has been active in the financial services sector.

Accion’s new $152.5M fund will back financial institutions serving small businesses globally

Meta’s newest social network, Threads is starting its own fact-checking program after piggybacking on Instagram and Facebook’s network for a few months. Instagram head Adam Mosseri noted that the company…

Threads finally starts its own fact-checking program

Looking Glass makes trippy-looking mixed-reality screens that make things look 3D without the need of special glasses. Today, it launches a pair of new displays, including a 16-inch mode that…

Looking Glass launches new 3D displays

Replacing Sutskever is Jakub Pachocki, OpenAI’s director of research.

Ilya Sutskever, OpenAI co-founder and longtime chief scientist, departs

Intuitive Machines made history when it became the first private company to land a spacecraft on the moon, so it makes sense to adapt that tech for Mars.

Intuitive Machines wants to help NASA return samples from Mars

As Google revamps itself for the AI era, offering AI overviews within its search results, the company is introducing a new way to filter for just text-based links. With the…

Google adds ‘Web’ search filter for showing old-school text links as AI rolls out

Blue Origin’s New Shepard rocket will take a crew to suborbital space for the first time in nearly two years later this month, the company announced on Tuesday.  The NS-25…

Blue Origin to resume crewed New Shepard launches on May 19

This will enable developers to use the on-device model to power their own AI features.

Google is building its Gemini Nano AI model into Chrome on the desktop

It ran 110 minutes, but Google managed to reference AI a whopping 121 times during Google I/O 2024 (by its own count). CEO Sundar Pichai referenced the figure to wrap…

Google mentioned ‘AI’ 120+ times during its I/O keynote

Firebase Genkit is an open source framework that enables developers to quickly build AI into new and existing applications.

Google launches Firebase Genkit, a new open source framework for building AI-powered apps

In the coming months, Google says it will open up the Gemini Nano model to more developers.

Patreon and Grammarly are already experimenting with Gemini Nano, says Google

As part of the update, Reddit also launched a dedicated AMA tab within the web post composer.

Reddit introduces new tools for ‘Ask Me Anything,’ its Q&A feature

Here are quick hits of the biggest news from the keynote as they are announced.

Google I/O 2024: Here’s everything Google just announced

LearnLM is already powering features across Google products, including in YouTube, Google’s Gemini apps, Google Search and Google Classroom.

LearnLM is Google’s new family of AI models for education

The official launch comes almost a year after YouTube began experimenting with AI-generated quizzes on its mobile app. 

Google is bringing AI-generated quizzes to academic videos on YouTube

Around 550 employees across autonomous vehicle company Motional have been laid off, according to information taken from WARN notice filings and sources at the company.  Earlier this week, TechCrunch reported…

Motional cut about 550 employees, around 40%, in recent restructuring, sources say

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: Watch all of the AI, Android reveals

Google Play has a new discovery feature for apps, new ways to acquire users, updates to Play Points, and other enhancements to developer-facing tools.

Google Play preps a new full-screen app discovery feature and adds more developer tools

Soon, Android users will be able to drag and drop AI-generated images directly into their Gmail, Google Messages and other apps.

Gemini on Android becomes more capable and works with Gmail, Messages, YouTube and more

Veo can capture different visual and cinematic styles, including shots of landscapes and timelapses, and make edits and adjustments to already-generated footage.

Google Veo, a serious swing at AI-generated video, debuts at Google I/O 2024

In addition to the body of the emails themselves, the feature will also be able to analyze attachments, like PDFs.

Gemini comes to Gmail to summarize, draft emails, and more

The summaries are created based on Gemini’s analysis of insights from Google Maps’ community of more than 300 million contributors.

Google is bringing Gemini capabilities to Google Maps Platform

Google says that over 100,000 developers already tried the service.

Project IDX, Google’s next-gen IDE, is now in open beta

The system effectively listens for “conversation patterns commonly associated with scams” in-real time. 

Google will use Gemini to detect scams during calls

The standard Gemma models were only available in 2 billion and 7 billion parameter versions, making this quite a step up.

Google announces Gemma 2, a 27B-parameter version of its open model, launching in June

This is a great example of a company using generative AI to open its software to more users.

Google TalkBack will use Gemini to describe images for blind people

Google’s Circle to Search feature will now be able to solve more complex problems across psychics and math word problems. 

Circle to Search is now a better homework helper

People can now search using a video they upload combined with a text query to get an AI overview of the answers they need.

Google experiments with using video to search, thanks to Gemini AI

A search results page based on generative AI as its ranking mechanism will have wide-reaching consequences for online publishers.

Google will soon start using GenAI to organize some search results pages

Google has built a custom Gemini model for search to combine real-time information, Google’s ranking, long context and multimodal features.

Google is adding more AI to its search results