Debatuj z Inteligencją – Gwiezdny Turniej Debat II

The proposition that artificial intelligence is the principal cause of rising disinformation has gained traction as generative models have become cheap, fast, and widely accessible. Tools capable of producing photorealistic images, cloned voices, and fluent text in any language emerged into mainstream use after the 2022 release of large language models such as ChatGPT and image generators like Midjourney and Stable Diffusion. By 2023-2024, documented deepfake incidents targeted elections in Slovakia, the United States, Bangladesh, and Indonesia, while AI-generated propaganda was tracked by researchers at NewsGuard, Graphika, and the Stanford Internet Observatory in operations linked to Russia (the 'Doppelganger' campaign) and China (the 'Spamouflage' network). The core dispute is causal: critics argue AI is the decisive accelerant of an information crisis, dramatically lowering the cost of producing convincing falsehoods and enabling micro-targeted manipulation at scale. Skeptics, including studies by the Alan Turing Institute and Meta's 2024 election review, counter that traditional drivers — partisan media ecosystems, social-platform recommendation algorithms, declining trust in institutions, and human willingness to share emotionally resonant content — remain the dominant factors, with AI amplifying rather than originating the problem. The debate now plays out against a thickening regulatory backdrop: the EU AI Act (2024) imposes transparency duties on generative systems, the U.S. issued Executive Order 14110, and UNESCO published its 2021 Recommendation on the Ethics of AI. Yet enforcement is uneven, watermarking standards (C2PA) are voluntary, and citizens across democracies report declining confidence in their ability to distinguish authentic from synthetic content — a perception that itself corrodes public discourse regardless of how much disinformation AI actually produces.

Artificial intelligence as a field dates to the 1956 Dartmouth Conference, but its current geopolitical salience stems from the deep-learning breakthroughs of the 2010s and the generative-AI inflection point of 2022-2023. Frontier models trained on vast datasets now perform at or above human level on many cognitive benchmarks, drive multi-trillion-dollar market valuations, and underpin applications from drug discovery to autonomous weapons. This rapid capability gain has outpaced national legal frameworks and triggered a global race between the United States and China for compute, talent, and standards leadership. The central policy disputes are threefold: how to manage catastrophic and systemic risks (bias, surveillance, autonomous weapons, biosecurity) without stifling innovation; how to distribute the economic gains given concentrations of compute and capital in a handful of firms and states; and how to govern AI across borders when training data, models, and harms all transcend jurisdiction. Export controls on advanced semiconductors imposed by the U.S. in 2022 and tightened since have made chips a frontline of strategic competition, while the Global South has pressed for inclusion in rule-making through forums such as the UN High-Level Advisory Body on AI. A patchwork of governance is emerging. The EU AI Act (in force 2024, with obligations phasing in through 2026) is the first comprehensive horizontal law; the Bletchley (2023), Seoul (2024), and Paris (2025) summits produced voluntary safety commitments from frontier developers; the UN General Assembly adopted its first AI resolution in March 2024; and the Council of Europe opened the first binding AI treaty for signature the same year. Yet fundamental questions — liability for model harms, treatment of open-weight systems, military uses, and the status of artificial general intelligence — remain unresolved.

Disinformation — the deliberate spread of false or misleading information — is not new, but its scale and velocity have been transformed by the internet, social platforms, and now generative AI. Modern concern crystallized after the 2016 U.S. presidential election and the Brexit referendum, when investigations exposed coordinated foreign influence operations, notably by Russia's Internet Research Agency. The COVID-19 pandemic produced what the WHO called an 'infodisemia,' and conflicts in Ukraine, Gaza, and the Sahel have demonstrated how falsified narratives shape battlefield outcomes and international opinion. The core dispute pits two values against each other: protecting citizens and democratic processes from manipulation, and safeguarding free expression from state or platform overreach. Authoritarian governments have invoked anti-disinformation laws to silence dissent (e.g., Russia's 2022 'fakes' law), while liberal democracies struggle with the line between content moderation and censorship. Platforms, once expected to self-police, have rolled back trust-and-safety teams since 2022, even as the EU's Digital Services Act and the UK's Online Safety Act impose new duties of care. Independent fact-checking networks and open-source investigators (Bellingcat, EU DisinfoLab) have become important counterweights. The international response is fragmented. UNESCO issued Guidelines for the Governance of Digital Platforms in 2023; the UN Secretary-General launched a Global Principles for Information Integrity initiative in 2024; the OECD and OSCE run dedicated programs. But there is no binding global instrument, definitions remain contested (disinformation vs. misinformation vs. malinformation), and the rise of synthetic media is forcing a shift from content-based regulation toward provenance, authentication, and media-literacy strategies.