Much of Aaron Sorkin’s new film “The Social Reckoning“ takes place in server rooms, where the algorithmic machinery that has shaped the last decade of human life is developed, maintained, and overseen. Sorkin’s focus is on “Meaningful Social Interactions,” or MSI, a ranking system that weights comments and reshares dozens of times — more heavily than likes. Thus, it supercharges rage, outrage, and lies. The film portrays MSI as a socio‑political accelerator, a system that rewards anger, punishes nuance, and destabilizes.
But today, there is a sense that its central crisis, the harms caused by engagement‑based ranking, is already being overshadowed by something even more powerful. The world we are in is ruled by advanced algorithms. The harms of MSI were a warm-up to generative Artificial Intelligence (AI) models being developed by extraordinarily rich, powerful tech companies, capable of producing text, images, audio, and video indistinguishable from human output. These systems do not merely curate reality; they manufacture it. And the stakes extend far beyond social cohesion or political polarization. They reach into the dangerous domains of fraudulent science and invented national security, thus upending planetary stability.
The Frontier Models and the New Era of Risk
While Sorkin’s film focuses on the algorithmic past, the present and future will be defined by frontier large‑scale generative models developed by OpenAI, Anthropic, Google DeepMind, and others. These systems represent a massive leap in capability. They can write persuasive political messaging tailored to individual psychological profiles. They can produce realistic synthetic video indistinguishable from authentic footage. They can simulate human conversation with emotional nuance. They can write code, design molecules, and optimize strategies. They can sway or destroy the minds of young people and adults (and, in some instances, are already doing so). They can operate autonomously across long sequences of tasks.
In other words, they are not merely tools; they are participants.
For the political and influencer universe, OpenAI’s GPT‑4 and Anthropic’s Claude 3 Opus can analyze vast amounts of information, generate coherent narratives, and adapt their tone and style to maximize impact. They can produce propaganda, misinformation, and manipulative content. A political campaign could use AI to generate millions of individualized messages, each crafted to exploit a specific voter’s emotional vulnerabilities. A foreign adversary could use AI to simulate thousands of synthetic activists, each pushing divisive narratives or simulated threats into local communities.
This is not entirely new. The Cambridge Analytica data scandal, which became public in 2018 and centered on the use of Facebook data in connection with the 2016 U.S. presidential election, offered an early and troubling illustration of how personal data and psychological profiling could be used for political targeting. The scandal became a watershed moment in the public debate over the power of digital platforms, revealing how information collected from millions of Facebook users could be used to build profiles for political advertising.
Frontier AI could take these practices further. Misinformation does not need humans — it can be generated automatically, endlessly, and convincingly. These systems can run simulations, test strategies, and refine their outputs based on feedback.
The Scientific and Security Risks
This has profound implications not only for politics and society, but for security and, in other words, the planet’s very stability. Such systems can accelerate or diminish scientific research, lower the barrier to biological experimentation, and provide step‑by‑step guidance for creating harmful agents. Researchers have already demonstrated that generative models can produce plausible protocols for synthesizing viruses, modifying pathogens, or bypassing laboratory safety systems. While responsible AI companies have implemented safeguards, the risk remains that future models — or unauthorized versions — could provide actionable guidance to individuals with malicious intent.
And the national‑security implications are alarming. Frontier AI could identify vulnerabilities in critical infrastructure, generate strategies to destabilize geopolitical rivals, or simulate escalation scenarios that increase the risk of conflict. In the most extreme cases, AI‑generated misinformation or fabricated military communications could trigger misunderstandings between nuclear‑armed states. A synthetic video of a missile launch, a fabricated audio clip of a military leader issuing orders, or a falsified intelligence report could create the conditions for catastrophic miscalculation.
Climate stability is also at risk because frontier AI systems could recommend actions that appear beneficial in the short term but have disastrous long‑term consequences. They could optimize economic efficiency in ways that accelerate environmental degradation. They could generate persuasive narratives that undermine climate science or encourage harmful geoengineering experiments.
Related Articles
Here is a list of articles selected by our Editorial Board that have gained significant interest from the public:
The Institutional Vacuum
One of the most striking elements of “The Social Reckoning” story is its portrayal of institutional failure — a failure basically the province of Mark Zuckerberg, co-founder of Facebook and its parent company, now named Meta. Facebook’s internal safety teams were understaffed, underfunded, and ignored. Executives prioritized bottom-line growth over responsibility. Regulators were outmatched by the speed and complexity of the platform.
But the institutional vacuum surrounding frontier AI is even more severe. There is not even a minimal regulatory regime governing the development or deployment of frontier AI. There are no binding safety standards, no mandatory audits, no enforceable transparency requirements. Companies like OpenAI and Anthropic operate in a largely self‑regulated environment. Anthropic’s President, Daniela Amodei, has said that “global health is becoming a focus for Anthropic,” an indication of its “effective altruism” goals, if it is to be believed to determine company policy.
Facebook had employees like Frances Haugen who leaked documents, but Frontier AI companies are even more secretive, more technical, and more centralized. The risks are harder to articulate, harder to prove, and harder to expose.
The public has little visibility into how frontier models are trained, what data they use, or how they behave under stress. The systems are proprietary, opaque, and controlled by a handful of companies. The decisions that shape the future of AI are being made by corporate CEOs, their boards, not elected officials. The public has no meaningful say in how these systems are deployed, what risks are acceptable, or what safeguards are necessary. Donald J. Trump, the President of the United States, has rejected federal legislation or statutory mandates for frontier AI companies, looking to industry to morally binding, voluntary controls.
The Reckoning Ahead
In other words, “The Social Reckoning” is a powerful indictment of the harm caused by engagement‑based ranking and the institutional failures — and greed — that allowed Facebook to flourish. This is only a shadow of what is to come: our world is one in which algorithms will shape not only behavior but reality itself.
Generative AI represents a fundamental shift in the nature of information. It is a tool that can empower creativity, accelerate innovation, and expand human potential. But it can also manipulate, deceive, and destroy. It has already reshaped society — the question is whether humanity can reshape and tame systems it has created. We must.
