The Pentagon Goes All-In on AI: Custom ChatGPT and Grok for “Every American Warrior”
Published: September 3, 2026 | By Vito Ruocco
From military AI platforms to LinkedIn’s war on slop, the AI landscape is shifting faster than ever. Here’s everything you need to know this week.
Introduction: The Week AI Changed the Rules of Engagement
This week marks a defining moment in the relationship between artificial intelligence and global power structures. In a series of announcements that would have seemed unthinkable just two years ago, the United States Department of Defense has launched a bespoke AI platform — GenAI.mil — bringing custom versions of ChatGPT, Grok, and Google’s Gemini directly to America’s fighting forces. Meanwhile, Google has quietly rewritten its AI ethics principles to remove long-standing commitments against developing AI for weapons and surveillance, signaling that the era of “responsible AI” guardrails is giving way to an arms race mentality.
These developments sit alongside a wave of fascinating stories: LinkedIn’s war on AI slop has exceeded a million user reports, OpenClaw 2.0 has landed with a radical reimagining of AI agency, and the legal system continues to grapple with AI hallucinations in high-stakes litigation. Let’s dive into the most consequential AI stories of the day.
1. GenAI.mil: The Pentagon’s New AI Arsenal
The Department of Defense surprised the world this week by announcing GenAI.mil, a generative AI platform purpose-built for the US military. The portal, which launched last year with Google’s Gemini, has now onboarded over 1.7 million personnel — more than half of the Pentagon’s 3 million-strong workforce.
Secretary of Defense Pete Hegseth — who has taken to calling himself “Secretary of War” — described the platform in characteristically dramatic terms: “It puts the world’s most powerful frontier AI models directly into the hands of every American warrior and will make our fighting force more lethal than ever before. The future of American warfare is here, and it’s spelled A-I.”
The platform now includes custom versions of OpenAI’s ChatGPT and xAI’s Grok, alongside Google’s Gemini, all “tailored to warfighter needs.” A Pentagon press release emphasizes these tools are for unclassified work only, covering tasks like:
- Summarizing policy handbooks and compliance documents
- Generating project-specific compliance checklists
- Extracting key terms from statements of work
- Creating detailed risk assessments for operational planning
Google was careful to note that data processed through GenAI.mil “is never used to train Google’s public models.” However, the announcement drew immediate concern from ethics watchdogs, especially given that Google reversed its commitment to avoid AI applications for weapons and surveillance earlier this year.
The rollout was apparently so rapid that at least one soldier took to Reddit, posting about a “weird pop up for the ‘Gen AI’ on my work computer” that “looks really suspicious.” The GenAI.mil portal itself is accessible at genai.mil, though it blocks unauthorized users with a stern access warning.
2. Google’s Quiet Ethical Retreat: AI Weapons Are Now on the Table
Perhaps no development underscores the shifting landscape more than Google’s quiet revision of its AI principles. Earlier this year, Google silently removed language that committed the company to not designing or deploying AI for weapons systems, surveillance technology, or applications “that cause or are likely to cause overall harm.”
The change, first spotted by The Washington Post, marks the end of a promise Google had maintained since 2018, when employees protested the company’s involvement in Project Maven — a Pentagon drone-imagery analysis program. At the time, Google pledged to never develop AI weapons. Now, that pledge is gone.
Demis Hassabis, CEO of Google DeepMind, and James Manyika published a blog post explaining the shift: “There’s a global competition taking place for AI leadership within an increasingly complex geopolitical landscape. We believe democracies should lead in AI development, guided by core values like freedom, equality, and respect for human rights.”
This aligns Google with competitors who already permit military use cases. Meta’s Llama and OpenAI’s ChatGPT already allow some forms of military use, while a partnership between Amazon, Palantir, and Anthropic brings Claude AI to US defense customers.
3. LinkedIn’s War on AI Slop Crosses a Milestone
On a lighter note, LinkedIn’s battle against AI-generated content reached a remarkable milestone this week. The platform’s “Seems like AI slop” reporting button — launched on July 30th — has now been clicked over 1 million times.
LinkedIn chief product officer Hari Srinivasan announced the milestone, revealing that users are “now experiencing 40% less views on what we classify as AI slop from just a few weeks ago.” The button, accessible from the three-dot menu on any post, is part of a broader crackdown that also includes new classifiers to detect AI-generated content and the removal of an “enhance your post” feature.
The push follows a Pangram analysis that found 41% of LinkedIn’s longform posts were flagged as fully AI-generated. LinkedIn is now also adding a new message that tells users when “some members told us this post seems like AI.”
In a stunning twist, even Microsoft’s own Office and LinkedIn chief is tired of AI slop. Ryan Roslansky, who oversees the Office org at Microsoft, recently posted a warning after receiving a fully AI-generated document from a colleague, calling out the “slop happening in the day to day of work.”
4. OpenClaw 2.0: The Open-Source Agent Revolution Arrives
OpenClaw, the open-source AI agent framework that has become a darling of the developer community, released version 2.0 this week — and it’s a monster. With 933 contributors (including 569 first-time contributors) and over 16,000 pull requests, founder Peter Steinberger described it as “accidentally” becoming the single largest update in the project’s history.
The release touches every component: simplified installation, a rebuilt browser app, improved messaging, memory, skills, model integration, automations, native apps, plugins, and security. But the headline feature is Shared Cloud Sessions, which turns OpenClaw into a multiplayer AI workspace where team members can collaborate on the same AI tasks in real-time.
Steinberger didn’t hold back: “Local harnesses feel like relics of the past now. Multiplayer coding + infinite compute kept everyone on the team up to date with others’ progress.”
The philosophy behind OpenClaw 2.0 is worth noting: rather than asking users to trust a single company or AI provider, OpenClaw remains fully open-source, letting users own their AI infrastructure. As the release notes state: “We are not selling anything here or asking you to trust one company, one model, or one AI provider with that future.”
5. Perplexity’s “Hybrid Compute” Brings Privacy to AI Agents
Perplexity quietly rolled out “Hybrid Compute” this week, a feature that splits AI processing between local and cloud-based models. Available on Perplexity Computer for Mac, the system uses a “privacy gate” to control which information leaves the device.
Cloud-based AI models handle reasoning, web search, and planning, while an on-device model processes tasks involving sensitive information and private files. This architecture addresses one of the biggest concerns in enterprise AI adoption: data privacy. For organizations handling sensitive documents, the ability to keep certain processing local while leveraging cloud-scale AI for complex reasoning is a significant step forward.
Perplexity Computer itself, launched earlier this year, describes itself as a “general-purpose digital worker” that “reasons, delegates, searches, builds, remembers, codes, and delivers.” It sits somewhere between OpenClaw and Anthropic’s Claude Cowork in the rapidly expanding AI agent ecosystem.
6. AI in the Courtroom: Hallucinations Strike Again
AI-generated errors continue to plague the legal system. This week, Mississippi state officials asked an appeals court to overturn a decision blocking restrictions on school DEI programs, arguing the ruling was “error-plagued” and drafted with the help of AI tool Perplexity. The dispute is the latest example of AI derailing court proceedings with made-up citations, bogus quotations, and dubious research.
In a separate incident, elite law firm Sullivan & Cromwell — which represents President Trump and handled the SpaceX/xAI merger — was forced to apologize to a federal judge for filing documents full of fake case citations hallucinated by AI. The list of errors ran three pages long.
These cases reinforce a growing consensus: language models are not actually intelligent. They generate plausible-sounding text without understanding truth, and when that text is fed into high-stakes legal proceedings, the consequences are severe. As Nilay Patel of The Verge put it: “Even the fancy lawyers are getting pantsed by AI.”
7. Artists Strike Back: Jason Isbell Sues Suno
Musician Jason Isbell filed a lawsuit against AI music company Suno, alongside Guy Forsyth, Eduardo Calle, and Camper Van Beethoven’s David Lowery. Rather than pursuing a straightforward copyright case, the plaintiffs claim Suno imitates their voices and styles without permission.
The lawsuit is notable for its framing: it’s not about specific copyrighted works being copied, but about the replication of artistic identity. The complaint argues that Suno’s AI models have been trained to mimic the distinctive vocal and stylistic signatures of these artists, effectively appropriating their creative identities without consent or compensation.
This case could set a crucial precedent for how AI music generation is regulated. If the courts recognize a right against AI voice/style imitation, it would force companies like Suno to either obtain licensing agreements or significantly alter their training approaches.
8. The “Artificial” Movie: Andrew Garfield as Sam Altman
In a fascinating convergence of life and art, Luca Guadagnino’s film “Artificial” — starring Andrew Garfield as OpenAI CEO Sam Altman — will have its world premiere at the New York Film Festival on October 5th.
The film chronicles the five-day drama of November 2023, when Altman was famously fired and then rehired by OpenAI’s board. Garfield’s casting has generated significant buzz, with early footage showing him capturing Altman’s distinctive blend of Silicon Valley earnestness and steely ambition.
The timing couldn’t be more appropriate: as AI reshapes warfare, workplaces, and creative industries, the story of the man at the center of the AI revolution is now entering the cultural bloodstream. “Artificial” promises to be a defining film for an era grappling with what it means when intelligence itself becomes a commodity.
What It All Means
Reading across these stories, a clear pattern emerges: AI is no longer a theoretical technology being debated in ethics papers and conference halls. It is being deployed in the military, in courtrooms, in music studios, on social media platforms, and inside every major software product.
The GenAI.mil launch represents perhaps the most significant inflection point. When the world’s most powerful military puts AI in the hands of 1.7 million personnel — with custom versions of the most advanced models available — the genie is well and truly out of the bottle. Google’s ethical retreat only confirms that the AI arms race is real, and major tech companies are choosing sides.
At the same time, the backlash against AI slop — whether on LinkedIn, in court filings, or in music — demonstrates a growing public awareness that AI-generated content requires human oversight, verification, and ethical boundaries. The million-plus LinkedIn users who clicked “Seems like AI slop” are sending a message: they want authentic human interaction, not algorithmic filler.
OpenClaw 2.0’s success, meanwhile, shows that there is a vibrant open-source alternative to the corporate AI agenda. When 933 contributors collaborate on a platform that lets users own their own AI infrastructure, it suggests that the future of AI might be more decentralized than the current narrative suggests.
One thing is certain: September 2026 will be remembered as the month when AI crossed from novelty into infrastructure. From the battlefield to the boardroom, from the courtroom to the recording studio, AI is now woven into the fabric of how we work, fight, create, and argue. The question is no longer whether AI will change everything — it’s whether we’re ready for what comes next.
About the author: Vito Ruocco is a technology analyst and writer covering artificial intelligence, cybersecurity, and the intersection of technology with society. Follow more analysis at ruocco.it.