Maria Ressa: How Big Tech and AI Have Undermined Democratic Life

Maria Ressa frames the problem as a progression that began more than a decade ago: companies designed platforms to capture attention, undermining journalism and civic information, and successive waves of AI intensified the harms. Her account links social-media business incentives to declines in public truth and describes three distinct AI phases that changed how people see, believe, and act.

Three phases of AI and their civic effects

Phase 1: social media and micro-targeting. Machine learning enabled personalized content that "cloned" people—platforms optimized for attention. That attention-economy business model eroded news organizations' revenue while re-engineering how publics consume information.

Phase 2: large language models (LLMs). Ressa emphasizes that LLMs are probabilistic machines and "have never been factual." Their public rollout in 2022–2023 as chatbots altered users' information environments: the outputs are not guaranteed truthful, and that shift changed expectations about facts, trust, and reliability.

Phase 3: agentic and multi-agent systems. The current stage moves beyond shaping attention and information to enabling software that can act on users' behalf. Ressa argues this targets agency itself: rather than merely consuming content that influences decisions, people increasingly give AI permission to perform tasks and make choices for them.

Evidence, reports, and incidents cited

Ressa points to the United Nations' Scientific Panel on Artificial Intelligence, which she co-chairs, and to a preliminary report released in July that warned the trends are "not controllable." She references specific model releases—including Mythos and Fable 5—and a public security incident involving Hugging Face as markers of how quickly capabilities and risks are evolving.

Ressa stresses that these problems did not begin overnight. Documented harms stretch back a decade, and limited policy action has followed. Present-day debates about guardrails for AI must account for the cumulative effect of attention-driven platforms plus successive AI rollouts that have degraded truth and shifted agency.

Ressa's roles and next steps for readers

The published transcript and a recorded FP Live conversation expand on these claims and pursue solutions. Ressa focuses public attention on the sequence of harms—attention capture, probabilistic language models, and agentic systems—and frames civic response as necessary because the tools reshape what people see and who decides.

• Policymakers and citizens should treat the harms as cumulative and evolving, not isolated events.
• Assessments of AI systems must account for their probabilistic nature and their capacity to perform actions for people.
• Face case are evidence points for urgent discussion about governance and public literacy.

Ressa's account is an argument for urgency: a decade of platform-driven changes plus successive AI advances have altered attention, truth, and agency. The remaining question—how citizens should respond—is the focus of her forthcoming book and the broader conversation she's advancing at the UN and in public forums.