The video conference that afternoon brought together the most diverse group of people Felix had ever seen on a single screen. Representatives from democratic AI governance networks in twelve countries filled the virtual meeting room, their faces reflecting the urgency and determination that had brought them together in response to the coordinated attacks.
"Before we continue," Emily said, glancing at her multiple time zone clocks, "I want to acknowledge we're asking Dr. Tanaka and Dr. Kim to join us at 2 AM their time, while Dr. Volkov has been up since 4 AM. This kind of global coordination means someone's always sacrificing sleep."
"And it's about to get worse," Felix added with a wry smile. "Half of us switch to daylight saving time next week, but on different dates. The US springs forward March 10th, Europe on March 31st, and some of you don't switch at all."
Dr. Yuki Tanaka from Tokyo laughed tiredly. "This is why I keep five clocks on my wall—Tokyo time, UTC, Eastern Standard, Eastern Daylight, and what I call 'American Chaos Time' for when I have to figure out if Arizona is matching California or not."
"Perhaps we should establish a rotating meeting schedule," Dr. Andersen suggested. "Each region takes a turn with the convenient time slot."
"Or we embrace asynchronous collaboration," Dr. Kim added, suppressing a yawn. "Not everything needs real-time discussion. We could use AI assistants to summarize discussions for different time zones—democratically governed AI helping us govern AI democratically."
The group agreed to implement both solutions: critical real-time meetings would rotate through time zones, while routine coordination would happen asynchronously with AI-generated summaries ensuring no region was left out of important decisions.
Dr. Tanaka spoke first, her voice carrying the measured precision of someone who had spent years building consensus in Japan's consensus-driven culture. "The attacks on our coordination network began at 3:17 AM local time," she said. "The pattern was identical to what you experienced in Pittsburgh—corrupted prompts designed to make our AI systems ignore worker welfare considerations."
Felix leaned forward in his chair, studying the faces on the screen. Each person represented not just a technical system, but a community of workers, companies, and citizens who had chosen to experiment with democratic governance of AI technology.
"The timing suggests global coordination," added Dr. Amara Okafor from Lagos, Nigeria. Her network had been one of the first to implement democratic AI governance in Africa, focusing on coordinating small-scale logistics operations across West Africa. "Our attack began exactly twelve hours after yours, which would be consistent with a follow-the-sun operation."
Dr. Lars Andersen from Copenhagen nodded grimly. "We've been analyzing the attack signatures," he said. "The techniques used are sophisticated enough to require significant resources and expertise. This isn't the work of individual hackers—it's a coordinated campaign by well-funded organizations."
Felix felt a mix of anger and determination as he listened to the reports. Each attack represented not just a technical failure, but an assault on the democratic values that these networks embodied. The attackers weren't just trying to break AI systems—they were trying to break the idea that ordinary people could participate in governing technology.
"What's our collective response?" asked Dr. Sarah Kim from Seoul. Her network had been particularly innovative in involving citizens in AI governance decisions, using deliberative polling and citizen juries to guide AI system development.
Emily Chen, who was moderating the call from her lab at Carnegie Mellon, pulled up a shared document on the screen. "I'm proposing that we formalize our cooperation through the Democratic AI Defense Initiative," she said. "This would be a global consortium focused on developing both technical and social defenses against attacks on democratic AI governance."
"The key principle," Emily continued, "is that we maintain the democratic character of our response. We don't become authoritarian just because we're fighting authoritarianism. We prove that democratic institutions can defend themselves while remaining democratic."
Dr. Raj Patel from Mumbai raised his hand. "What would this look like in practice?" he asked. His network had been working on coordinating supply chains for small manufacturers across India, with a particular focus on ensuring that AI optimization didn't disadvantage rural producers.
Emily clicked to a new slide showing the proposed structure. "We would have four working groups," she said. "Technical Defense, focused on developing robust AI systems that can resist adversarial attacks. Policy Response, focused on working with governments to develop appropriate regulatory frameworks. Community Organizing, focused on building public support for democratic AI governance. And International Coordination, focused on sharing information and resources across networks."
"Each working group would include representatives from multiple networks and countries," she continued. "Decisions would be made through consensus-building processes that respect cultural differences while maintaining shared commitments to democratic values."
Dr. Elena Volkov from Moscow leaned forward. "What about networks in countries where the government might not support democratic AI governance?" she asked. Her network operated in a challenging political environment, focusing on coordinating humanitarian aid distribution while avoiding government interference.
"That's exactly why international coordination is so important," Felix said. "Networks in more supportive political environments can provide resources and protection for networks in more challenging situations."
Dr. Tanaka nodded. "In Japan, we have strong government support for AI research and development. We could potentially provide funding and technical resources for networks in countries where such support isn't available."
"And in Nigeria," Dr. Okafor added, "we have strong community support and innovative approaches to participatory governance. We could share our methods for involving citizens in AI decision-making."
Felix felt the energy in the virtual room building as people began to see the potential for genuine collaboration. "This is exactly what the attackers don't want," he said. "They want us isolated and defensive. Instead, we're building a global movement."
Dr. Kim from Seoul raised a concern. "How do we ensure that this initiative doesn't become dominated by networks from wealthy countries?" she asked. "Democratic AI governance needs to work for everyone, not just those with the most resources."
Emily nodded. "That's a crucial point," she said. "The governance structure needs to ensure equal representation regardless of the size or resources of individual networks. We're proposing a rotating leadership structure where each region takes turns leading different working groups."
"We also need to ensure that the technical solutions we develop are accessible to networks with limited resources," added Dr. Andersen from Copenhagen. "Open-source development and knowledge sharing need to be core principles."
Dr. Patel from Mumbai smiled. "In India, we have a saying: 'Many hands make light work.' If we pool our resources and expertise, we can develop solutions that none of us could create alone."
Dr. Volkov suddenly unmuted with urgency in her voice. "Wait—before we continue, I need clarification. When you say 'open-source,' do you mean completely public? Because in Russia, if our defensive algorithms become public knowledge, the FSB will immediately classify them as dual-use technology. We could face criminal prosecution."
An uncomfortable silence fell over the call. Dr. Kim from Seoul shifted uncomfortably. "We have similar concerns in South Korea. Our military considers AI defense systems to be national security assets."
"This is exactly the kind of challenge we need to address," Emily said, thinking quickly. "We can create a tiered sharing system—some components fully open-source for basic protection, others shared only among trusted network members under specific security protocols."
Dr. Tanaka suggested, "In Japan, we've developed legal frameworks for 'controlled open-source'—transparent to participants but protected from adversarial access. We could adapt this model."
"But who decides what gets shared at which level?" Dr. Okafor asked pointedly. "We can't have wealthy nations keeping the best defenses for themselves while giving us in Africa only basic protection."
Felix recognized the tension building. "You're right, Amara. Equal access to defense is non-negotiable. How about this—each network contributes what they can legally share, but everyone gets access to the full defensive capability through secure implementation support?"
After several minutes of heated discussion, they reached a compromise: critical defensive capabilities would be shared through secure channels with legal protections, while general principles and non-sensitive components would be fully open-source.
The conversation continued for over two hours, with participants sharing their experiences, challenges, and innovations. Felix was struck by the diversity of approaches represented on the call. Each network had adapted democratic AI governance principles to their local context, creating a rich ecosystem of experimentation and learning.
Dr. Okafor from Lagos had developed innovative methods for involving informal sector workers in AI governance decisions. Dr. Volkov from Moscow had created sophisticated techniques for protecting democratic processes from government interference. Dr. Kim from Seoul had pioneered the use of digital platforms for large-scale citizen participation in AI policy-making.
"What we're seeing," Emily observed, "is that democratic AI governance isn't a single model that gets replicated everywhere. It's a set of principles that get adapted to local conditions and cultures."
"That's actually our strength," Felix added. "The attackers are trying to prove that democratic AI governance is inherently flawed. But what they're really attacking is a diverse ecosystem of democratic innovations. Even if they succeed in compromising some networks, others will survive and adapt."
Dr. Tanaka raised a strategic point. "We need to think about how to go on the offensive," she said. "Right now, we're responding to attacks. But we should also be demonstrating the positive benefits of democratic AI governance."
"What do you mean?" Dr. Andersen asked.
"We should be showing that democratic AI governance doesn't just defend against attacks—it produces better outcomes than corporate-controlled alternatives," Dr. Tanaka explained. "Better outcomes for workers, better outcomes for communities, better outcomes for society as a whole."
Felix felt excitement building. "That's exactly right," he said. "We need to shift the narrative from 'Can democratic AI governance defend itself?' to 'Why would anyone choose corporate-controlled AI when democratic alternatives work better?'"
Emily pulled up a new slide showing comparative performance data. "We actually have evidence for this," she said. "Networks using democratic AI governance consistently show better outcomes on measures of worker satisfaction, community benefit, and long-term sustainability."
"The problem is that this evidence isn't widely known," Dr. Kim observed. "We need better communication strategies to make sure policymakers and the public understand the benefits of democratic AI governance."
"That's where the Community Organizing working group comes in," Felix said. "We need to build public support for democratic AI governance by demonstrating its benefits in concrete, tangible ways."
Dr. Okafor leaned forward. "In Nigeria, we've found that the most effective way to build support is to involve people directly in the governance process," she said. "When people see that they can actually influence how AI systems affect their lives, they become strong advocates for democratic governance."
"That's a key insight," Emily said. "Democratic AI governance isn't just about better technical systems—it's about empowering people to participate in decisions that affect their lives."
The conversation turned to practical next steps. Each network committed to contributing resources to the Democratic AI Defense Initiative, whether in the form of funding, technical expertise, or community organizing capacity.
"We need to talk about sustainable funding," Dr. Andersen said pragmatically. "Voluntary contributions won't be enough. We need a real economic model."
Emily pulled up a funding proposal. "We're proposing a three-tier approach. First, participating networks contribute 0.5% of their operational savings from democratic governance—our data shows most networks save 15-20% compared to traditional systems, so this is affordable. Second, we're applying for grants from the EU, UN Development Programme, and several foundations interested in democratic technology. Third, we're exploring a mutual insurance model where networks pool resources to cover losses from successful attacks."
"The insurance model is brilliant," Dr. Patel said, his business background showing. "In India, we have similar cooperative insurance systems for farmers. Members have incentive to help each other improve security because everyone's premiums depend on collective resilience."
Dr. Kim added, "Samsung and Hyundai have expressed interest in corporate sponsorship if we can demonstrate measurable security improvements. They see democratic AI governance as a competitive advantage and want to support the infrastructure."
"Just be careful about corporate capture," Dr. Volkov warned. "The moment we become dependent on corporate funding, we lose our independence."
Felix nodded. "Agreed. That's why we need diverse funding sources with no single contributor controlling more than 10% of our budget."
"We need to move quickly," Dr. Volkov warned. "The attackers won't wait for us to get organized. They'll continue to escalate their campaigns while we're building our defenses."
"Then we work in parallel," Felix said. "We implement immediate defensive measures while building long-term collaborative capacity."
Emily outlined a timeline for the next phase. "Within two weeks, we'll have the first version of our adversarial detection system deployed across all participating networks. Within a month, we'll have enhanced constitutional training methods that are more resistant to value learning attacks. Within three months, we'll have a comprehensive defense framework that can be adapted to any democratic AI governance network."
"How exactly will networks with completely different architectures integrate these defenses?" Dr. Okafor asked. "Our system in Lagos runs on distributed mobile nodes with limited computing power. Dr. Kim's network in Seoul uses quantum-enhanced processing. These aren't compatible."
Emily brought up a technical diagram. "We're developing what we call 'defensive middleware'—a translation layer that adapts protection mechanisms to different architectures. Think of it like how USB-C adapters let different devices connect. The core defensive algorithms remain the same, but the implementation layer translates them for each network's specific architecture."
"For example," she continued, "Seoul's quantum systems can run full homomorphic encryption for maximum security, while Lagos's mobile nodes might use lightweight hash-based signatures. Both achieve protection, just through different means appropriate to their infrastructure."
Dr. Patel leaned forward. "We faced similar challenges connecting modern systems with legacy infrastructure in rural India. We created modular defenses—basic modules everyone can run, advanced modules for those with more resources, but all interoperable."
"The key is standardizing the threat intelligence sharing protocol," Emily explained. "Even if networks can't run identical defenses, they can share attack signatures in a common format. When Seoul's quantum systems detect a new attack pattern, they can immediately share that intelligence with Lagos's mobile network, which implements appropriate defenses for its architecture."
"And throughout this process," Emily continued, "we'll be documenting everything and sharing it openly. The attackers operate in secret, but we operate in the open. That's both our vulnerability and our strength."
As the call wound down, Felix felt a sense of hope that had been missing since the attacks began. The attackers had succeeded in compromising individual networks, but they had also catalyzed the creation of a global movement for democratic AI governance.
"There's one more thing," Dr. Tanaka said as the meeting was ending. "We need to remember that this isn't just about defending our current systems. It's about building the future we want to see."
"What do you mean?" Dr. Patel asked.
"The attackers represent a vision of the future where AI technology is controlled by a small number of powerful corporations and governments," Dr. Tanaka explained. "We represent a different vision—a future where AI technology serves democratic values and empowers ordinary people."
"This conflict isn't just about cybersecurity," she continued. "It's about what kind of society we want to live in. Do we want a future where technology concentrates power in the hands of a few, or do we want a future where technology distributes power more broadly?"
Felix nodded, feeling the weight and importance of the choice they were making. "Then let's build the future we want to see," he said. "Let's prove that democratic AI governance isn't just possible—it's better."
As the video conference ended and the screens went dark, Felix remained in Emily's lab, thinking about the global alliance they had just formed. The attackers had resources, expertise, and powerful allies. But the democratic AI governance movement had something more important—it had the support of people around the world who believed that technology should serve human welfare rather than concentrated power.
The war for the future of AI was far from over, but for the first time since the attacks began, Felix felt confident that democracy would prevail.