The Theater of “Responsible”AI Leadership

What Anthropic and OpenAI Are Really Saying About AI Governance

In the span of 48 hours this past week, the CEOs of two of the world’s most powerful AI companies published sweeping statements about the future of their technology. Dario Amodei of Anthropic released “When AI Builds Itself,” a detailed technical analysis of recursive self-improvement. Sam Altman of OpenAI, co-authoring with Chief Scientist Jakub Pachocki, published “Built to Benefit Everyone: Our Plan,” a broad vision statement about access, safety, and shared prosperity.

Read together, the two pieces paint a portrait of an industry that wants you to believe it is grappling seriously with the consequences of what it has built. But the question worth asking is whether these are genuine calls for oversight, or whether they are something more calculated: public positioning designed to create the appearance of self-regulation so that actual government regulation never arrives.

What They Are Saying

Anthropic’s piece is the more technically grounded of the two. Using internal data from its own engineering teams, the company reveals that over 80% of the code merged into Anthropic’s codebase is now written by Claude, its AI system. Engineers are shipping eight times as much code per quarter as they were in 2024. Claude’s success rate on open-ended engineering tasks has climbed 50 percentage points in six months. The piece lays out three possible futures: one where progress stalls, one where humans continue to set research direction while AI handles execution, and one where AI systems become capable of fully designing their own successors, a scenario Anthropic calls “recursive self-improvement.”

The company says it would welcome the option to slow or temporarily pause frontier AI development. It calls for international coordination mechanisms, verification systems, and shared safety standards among labs. It explicitly states that if such systems existed, Anthropic would slow down, as long as other frontier developers did so in a verifiable manner.

OpenAI’s piece is vaguer and more aspirational. Altman and Pachocki compare AI to electrification and speaks about how AI can help humanity like electricity did. The authors argue that the technology should be available to everyone, and outline three goals: building an automated AI researcher, accelerating the economy, and giving every person on Earth a “personal AGI.” The piece calls for international coordination and suggests there should ultimately be an organization that helps leading AI efforts reduce catastrophic risk. It also says the human role becomes more important as systems become more capable, emphasizing that “entirely automating everything is not the future we want.”

Both pieces express a preference for broad distribution of power rather than concentration. Both acknowledge risks. Both gesture toward the need for some form of governance.

The “I Will If You Go First” Problem

Look closely at Anthropic’s framing and you will notice something important. The company says it would slow down, if such governance systems existed, and if other frontier developers also slowed down in a verifiable manner. That is a conditional offer resting on two things that do not exist and that no one is building with any urgency or even at all.

This is the AI industry’s version of “I will if you do it first.” No company wants to be the one that voluntarily slows development, because slowing down means losing market share, losing talent, losing investor confidence, and potentially losing the technological lead entirely. Both Anthropic and OpenAI frame the competitive dynamic the same way the Cold War framed nuclear weapons: whoever wins the AI race controls the future, and falling behind is an existential risk. Anthropic’s piece explicitly argues that a unilateral pause by one lab would simply change who the front-runner is without creating any broader deliberative process.

The logic is understandable. It is also self-serving. It allows every company to keep building at maximum speed while pointing at the others and saying, “We cannot stop because they will not stop.” It is a perfectly circular justification for inaction, and every player at the table benefits from it.

Here is the uncomfortable truth: coordination is not going to come from these companies. It never has in the history of industry. It will have to come from government oversight, either at the level of individual nations or through an international commission with real authority to set and enforce boundaries on AI development and use. Something like the EU Commission, but broader in scope and with binding enforcement power across borders.

History Already Answered This Question

Both essays treat the question of governance as though it is novel, as though humanity has never faced the challenge of regulating a powerful technology controlled by a small number of companies with enormous financial incentives to resist constraint. We have faced it many times. The answer has been the same every time.

During the Industrial Revolution, factory owners did not voluntarily limit child labor, shorten working hours, or install safety equipment. They did those things when governments passed laws requiring them to. The reason was not that factory owners were uniquely evil. The reason was that the competitive dynamics made voluntary action irrational. Any single factory that raised its costs by treating workers better would be undercut by its competitors. Only an external authority applying the same rules to everyone could break the cycle.

The breakup of Ma Bell followed the same pattern. AT&T had become a monopoly that controlled the entire telecommunications infrastructure of the United States. The company argued, just as AI companies argue now, that its dominance was necessary for the system to work properly, that competition would create chaos and fragmentation, that it was the responsible steward. The government broke it up anyway, and the result was an explosion of innovation and competition that gave us the modern telecommunications industry.

The same story played out with Standard Oil, with the railroads, with the tobacco industry, with financial services after 2008. At every point in our collective history, successful companies that controlled transformative technologies had to be regulated to prevent self-dealing, harm to workers, harm to customers, and harm to the environment. The companies always argued they could govern themselves. They never could. Not because the people running them were bad people, but because human nature does not change based on the size of the technology. The incentives to prioritize profit, growth, and competitive advantage over public welfare are structural. They are baked into the way markets work.

Corporate law in the United States is designed around this exact understanding. The rules governing officers and directors, fiduciary duties, prohibitions on self-dealing, conflict of interest disclosures, these exist because we learned the hard way that humans in charge of companies cannot be trusted to refrain from self-dealing without harsh consequences for doing so. That is not cynicism. It is the foundational insight of modern corporate governance.

Why would we think humans in charge of AI technology would be any different? Why would we assume that the people building the most powerful and profitable technology in human history will choose to slow down, share power, and accept constraints voluntarily, when no comparable group of humans has ever done so before?

What They Are Not Saying

Conspicuously absent from both statements is any reckoning with what their technology is already doing right now.

The Jobs Are Already Being Lost

The AI industry talks about the future in hypotheticals while the present is already brutal. In the first five months of 2026 alone, American tech companies eliminated more than 142,000 jobs, a 33% increase over the same period the year before. And these cuts are happening at companies posting record revenues. Over 55% of layoff events in 2026 explicitly cite AI, automation, or machine learning as a contributing factor, impacting roughly 152,000 workers. Block, the fintech company behind Square and Cash App, cut its workforce from 10,000 to fewer than 6,000 in a single round, with CEO Jack Dorsey saying openly that the reductions were driven by AI capability, not financial difficulty.

Anthropic’s own CEO, Dario Amodei, has publicly stated that AI will wipe out half of entry-level white-collar jobs in the U.S. A Stanford study found that software developer employment for workers under 26 has fallen nearly 20% since 2024. An MIT simulation showed AI can replace nearly 12% of the entire U.S. workforce, amounting to roughly $1.2 trillion in lost wages.

Altman’s piece talks about giving everyone a “personal AGI” and “ensuring the gains are widely shared.” But his company just signed a Pentagon contract within hours of its chief rival being frozen out. When he says “everyone,” who exactly does he mean? The workers being laid off to fund the $700 billion data center buildout his company is part of? The four hyperscalers alone, Amazon, Microsoft, Alphabet, and Meta, have committed to a combined $700 billion in capital expenditure for 2026, nearly double what they spent in 2025. That money comes from somewhere. Quite a lot of it comes from cutting the humans who used to do the work.

The Environmental Cost Is Staggering

Neither piece mentions the environment. Not once.

A United Nations University report released in June 2026 found that global data centers used 448 trillion watt-hours of electricity in 2025, more than all but ten countries on Earth. That energy use produced roughly 208 million tons of carbon dioxide, about the same as Argentina, and consumed approximately 1.2 trillion gallons of water. The report predicts data center water and energy use will double in the next four years as AI demand grows.

More than 230 environmental organizations have collectively called on Congress to place a national moratorium on new data center construction. Lawmakers in over 30 states have introduced more than 300 bills addressing data center impacts in 2026 alone. AI training clusters consume seven to eight times more energy than typical computing workloads, according to MIT researchers.

Anthropic talks about the possibility that compute supply might constrain AI progress. It frames this as a potential bottleneck to development. It does not frame it as communities losing access to affordable electricity and clean water so that AI companies can train their next model.

The Surveillance and Weapons Pipeline Is Already Built

Perhaps the most revealing contrast between what these companies say and what they do lies in the military space.

Anthropic deserves credit for a specific act of refusal. In February 2026, the Pentagon demanded unrestricted access to Claude for “all lawful purposes.” Amodei required two exceptions: no mass surveillance of Americans and no fully autonomous weapons. The Pentagon refused. Defense Secretary Pete Hegseth gave Anthropic a three-day ultimatum. When the company held its ground, the Trump administration ordered federal agencies to stop using Anthropic’s products and designated the company a supply chain risk, a label that had never before been applied to an American company. Amodei called the response “very hard to interpret in any way other than punitive.”

What happened next tells you everything you need to know about OpenAI’s version of “safety principles.” Within hours of Anthropic being frozen out, Sam Altman announced that OpenAI had signed a deal with the Pentagon to deploy its models on classified military networks. In an internal note to staff, Altman said OpenAI would seek to negotiate exclusions preventing use for surveillance in the U.S. or to power autonomous weapons without human approval. But as The Intercept reported, there is no publicly available proof that any of these exclusions were actually enshrined in the contract. OpenAI’s legal framing “references laws,” meaning it defers to whatever the government decides is legal, rather than drawing its own lines.

Altman later acknowledged the deal was “rushed” and “looked opportunistic and sloppy.” That may be the most honest thing in either of these pieces.

Meanwhile, AI is already deeply embedded in the surveillance and weapons apparatus regardless of what either company does. Palantir holds a $10 billion Army Enterprise Agreement and a $30 million ICE contract to build what is called ImmigrationOS. Palantir’s Gotham software, originally designed to predict IED attacks in Afghanistan, is now used by hundreds of police departments across the United States to scrape datasets of license plate records, utility bills, and social media to build intelligence dossiers on civilians. Its ELITE tool reportedly mines Medicaid and other public welfare data for immigration enforcement. Technologies perfected on battlefields are being repackaged for domestic law enforcement, and AI has made that pipeline faster, cheaper, and harder to detect.

The timing of recent announcements makes the surveillance supply chain even more concerning. On June 4, 2026, Palantir announced a sweeping partnership with Google Cloud at its annual AIPCon event. The deal establishes two-way data pipelines between Palantir’s Foundry platform and Google’s BigQuery, links Palantir’s Ontology system with Google’s Knowledge Catalog, and, most critically, creates direct integration between Google’s Gemini AI models and Palantir’s AIP platform. Four days later, at WWDC 2026, Apple confirmed what had been rumored for months: the rebuilt Siri AI is powered by a custom version of Google’s Gemini, under a multi-year deal reportedly worth $1 billion a year. Siri AI will run across every Apple device in the ecosystem, including iPhone, iPad, Mac, Apple Watch, AirPods, Vision Pro, and CarPlay in your car. Follow the thread: Palantir, the company that builds surveillance dossiers on civilians for police departments and mines public welfare data for immigration enforcement, now has a deep technical integration with Google’s Gemini. Google’s Gemini now powers the voice assistant that will live on every Apple device on the planet, listening for commands in your living room, reading your messages, analyzing your emails, and riding along in your vehicle. Nobody signed a single consent form connecting those dots. Apple frames this as a privacy-first architecture. Google calls Palantir a partner in “AI-driven insights.” Palantir calls its police surveillance tool “Gotham.” The infrastructure that connects a defense contractor’s targeting platform to the AI running on your phone already exists. Whether anyone chooses to use it that way is a policy question, and right now, there is no policy. And these articles by Amodei and Altman make it clear they don’t want a policy in place, ever.

The Governance Problem Is a Conflict of Interest Problem

If an international AI governance body is ever created, and it needs to be, the single most important design question will be who sits on it.

The people and institutions designing and implementing governance frameworks cannot be invested in AI companies or in the broader tech ecosystem that profits from AI expansion. This sounds obvious, but it is exactly the kind of principle that gets quietly abandoned when powerful industries have a seat at the table. We see it in financial regulation, where former bank executives staff the agencies that oversee their old employers. We see it in pharmaceutical regulation, where advisory panels include researchers funded by the companies whose drugs they are evaluating. We see it in environmental regulation, where industry lobbyists write the rules their clients are supposed to follow.

Conflict of interest is not an abstract legal concept. It is a recognition of human nature. People who stand to profit from a particular outcome cannot be trusted to regulate that outcome fairly. Not because they are corrupt in some cartoon villain sense, but because self-interest distorts judgment in ways that are often invisible to the person experiencing it. The entire edifice of corporate governance law, fiduciary duty, duty of loyalty, the business judgment rule, prohibitions on self-dealing, exists because centuries of experience taught us that good intentions are not enough. You need rules, transparency, and consequences.

Any serious AI governance framework has to take this reality seriously. Commissioners cannot hold stock in AI companies. They cannot accept consulting fees or advisory roles from the industry they regulate. Their staff cannot cycle in and out of AI company employment the way defense officials currently cycle through defense contractors. The revolving door that brought former Anduril executives and Pentagon officials into OpenAI’s leadership, which then helped the company secure military contracts, is exactly the kind of structural conflict that governance frameworks need to prevent.

Regulation or Posturing?

So back to the central question: are these companies genuinely open to regulation, or is this public performance?

Anthropic’s Amodei walked away from a Pentagon contract, took a direct financial and political hit, and is suing the federal government. That is not posturing because the company is now facing real consequences for standing on principle. The “When AI Builds Itself” piece, with its internal data and candid employee quotes about feeling obsolete, reads as a genuine attempt to be transparent about a trajectory the company itself finds concerning. But the conflict of interest is apparent when you read it carefully. While Amodei is admitting the risks, he is also saying that he won’t take the lead in advocating for regulatory oversight.

Anthropic is an AI company racing to build more powerful systems. The piece carefully frames “recursive self-improvement” as something that “could come sooner than most institutions are prepared for” while continuing to develop the very technology that makes it possible. The call for international coordination and verifiable slowdowns sounds responsible, but it is a call for someone else to take the lead (and the hit) in creating a structure that does not currently exist. Without someone being willing to step forward and do the courageous thing, we will not have any slow down. So while sounding “transparent”, Amodei’s position conveniently allows full-speed development to continue. The conditional nature of the offer, “we will slow down if the governance mechanisms exist and if others also slow down,” functions as permission to keep going. It is a promise contingent on conditions that no AI CEO is willing to create.

OpenAI’s piece is harder to take at face value. Altman compares AI to electrification and talks about broadly distributed power while his company actively pursued and secured a Pentagon contract the moment its rival was pushed aside. The piece says “entirely automating everything is not the future we want” while the company’s own trajectory, mirroring Anthropic’s internal data, shows that precisely this kind of automation is underway. Altman’s piece calls for international coordination “including slowing frontier development when needed,” but OpenAI has shown no indication that it would actually slow anything and Altman does not define “when needed.” Presumably, he doesn’t believe it will ever be needed. The company removed its prohibition on military use from its policies in January 2024. It has staffed up with defense industry veterans. It announced a Pentagon deal within hours of Anthropic’s refusal. It has filed a confidential S-1 with the SEC. Every action the company takes points toward acceleration, not restraint.

The pattern is familiar from every other transformative industry in history. Tobacco companies talked about responsible marketing while expanding into new markets. Oil companies published sustainability reports while lobbying against emissions regulations. Financial institutions promoted self-regulatory frameworks while building the derivative structures that crashed the global economy. Human psychology and the playbook have not changed: speak the language of responsibility loudly enough that legislators believe the industry can govern itself, then continue operating without meaningful constraint. And contribute generously to lawmaker’s campaigns and pet projects to ensure no constraint will ever be seriously contemplated or enacted.

What Would Genuine Accountability Look Like?

If these companies were serious about the concerns they raise in their own publications, we would see them actively supporting specific legislation rather than calling for “conversations.” We would see them publishing their military contracts in full rather than issuing vague assurances about “technical safeguards.” We would see environmental impact statements for every new training run. We would see them funding retraining programs and transition support for the workers their technology is displacing rather than talking about a future where “the gains are widely shared.” We would see them advocating for a governance body with real enforcement power and insisting that its members have no financial ties to the industry.

We would also see them taking the lead on addressing the economic displacement their technology is creating. Multiple AI leaders, including Anthropic’s own CEO, have said publicly that mass job loss is inevitable. And yet none of them are working to ensure that displaced workers can still survive. They are not funding retraining at scale. They are not proposing new economic models for a world where human labor is no longer competitive. They are focused on their own net worth, access to powerful people and government contracts, and racing toward massive IPOs, which is exactly what OpenAI was doing the same day Altman published his essay about benefiting all of humanity.

The environmental reckoning is equally absent. The energy and water consumption of data centers is already rivaling that of entire nations, and it is projected to double within four years. The mineral extraction required to build the hardware, the satellite infrastructure polluting low-Earth orbit, the strain on local power grids and water systems, none of this appears in their vision statements. They are building a future that may not be physically sustainable, and they are not acknowledging it.

What they are doing, quietly, is preparing for the worst. Many of the same tech leaders writing essays about shared prosperity are also purchasing remote properties, building private bunkers, and investing in personal survival infrastructure. If they genuinely believed the future they describe in their publications, they would not be hedging against civilizational collapse. The bunkers tell you what they actually think is coming. The essays are for everyone else, and they serve a second purpose: buying time. Every month spent on “conversations” and “international coordination frameworks” is another month to consolidate wealth, close government contracts, and launch IPOs before the political winds shift. The November midterms are approaching, and public anger over AI-driven job losses, rising energy costs, and unchecked surveillance is building. These CEOs are not writing essays about responsibility because they plan to slow down. They are writing them because they need the window to stay open long enough to finish grabbing everything they can before the window closes.

Instead, we get beautifully written essays about the importance of human judgment, published by companies that are systematically automating human judgment out of existence. We get calls for “distributed power” from organizations concentrating unprecedented capability in their own hands. We get warnings about recursive self-improvement from the people building it. We get conditional offers to slow down that are structured so the conditions will never be met.

These essays are not useless. They contain real data, real concerns, and in Anthropic’s case, real evidence of a company willing to pay a price for a principle. But they are also insufficient. The gap between what these companies say and what is actually happening, the layoffs, the environmental damage, the surveillance infrastructure, the weapons integration, is enormous.

The question for readers, policymakers, and the public is whether we are going to wait for the AI industry to regulate itself, or whether we are going to demand that democratic institutions do the job they exist to do. History has already answered this question. We just have to be willing to listen.

Sources:

The Two Essays

• Anthropic, “When AI Builds Itself”: https://www.anthropic.com/institute/recursive-self-improvement
• OpenAI, “Built to Benefit Everyone: Our Plan”: https://openai.com/index/built-to-benefit-everyone-our-plan/

AI Layoffs and Job Displacement

• Skillsyncer 2026 Tech Layoffs Tracker: https://skillsyncer.com/layoffs-tracker
• Tech Insider, “Tech Layoffs 2026: How AI Is Driving the Biggest Workforce Shift”: https://tech-insider.org/tech-layoffs-2026-ai-workforce-impact/
• TechTimes, “Tech Layoffs Reach 142,000 in 2026”: https://www.techtimes.com/articles/317392/20260529/tech-layoffs-reach-142000-2026-profitable-companies-cut-jobs-fund-700b-ai-infrastructure.htm
• Tom’s Hardware, “Tech Industry Lays Off Nearly 80,000 in Q1 2026”: https://www.tomshardware.com/tech-industry/tech-industry-lays-off-nearly-80-000-employees-in-the-first-quarter-of-2026-almost-50-percent-of-affected-positions-cut-due-to-ai
• Final Round AI, “AI and Tech Layoffs in 2025-2026”: https://www.finalroundai.com/blog/tech-layoffs-ai-2026
• Programs.com, List of Companies Announcing AI-Driven Layoffs: https://programs.com/resources/ai-layoffs/

Environmental Impact

• Washington Times/AP, “Energy, Water Use and Pollution of AI and Data Centers Rival Most Countries” (UN University Report): https://www.washingtontimes.com/news/2026/jun/3/report-energy-water-use-pollution-ai-data-centers-rival-countries/
• ANSI Blog, “Making AI Data Centers More Sustainable”: https://blog.ansi.org/ansi/ai-data-centers-carbon-water-energy-impact/
• Consumer Reports, “AI Data Centers: Big Tech’s Impact on Electric Bills, Water, and More”: https://www.consumerreports.org/data-centers/ai-data-centers-impact-on-electric-bills-water-and-more-a1040338678/
• MIT News, “Explained: Generative AI’s Environmental Impact”: https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117
• NIH/Patterns Journal, “The Carbon and Water Footprints of Data Centers”: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12827721/

Anthropic-Pentagon Dispute

• NPR, “OpenAI Announces Pentagon Deal After Trump Bans Anthropic”: https://www.npr.org/2026/02/27/nx-s1-5729118/trump-anthropic-pentagon-openai-ai-weapons-ban
• CBS News, “AI Executive Dario Amodei on the Red Lines Anthropic Would Not Cross”: https://www.cbsnews.com/news/ai-executive-dario-amodei-on-the-red-lines-anthropic-would-not-cross/
• CBS News, “Anthropic CEO Says He’s Sticking to AI Red Lines”: https://www.cbsnews.com/news/pentagon-anthropic-dario-amodei-cbs-news-interview-exclusive/
• Newsweek, “Anthropic Chief Reacts to Pentagon Feud”: https://www.newsweek.com/anthropic-chief-reacts-to-pentagon-feud-red-lines-11597807
• Fortune, “Anthropic CEO Dario Amodei Says ‘We Are Patriotic Americans’”: https://www.fortune.com/2026/02/28/anthropic-ceo-dario-amodei-patriotic-americans-trump-hegseth-mass-surveillance-autonomous-weapons
• PPC Land, “Anthropic’s CEO Drew Two Red Lines. The Pentagon Said No.”: https://ppc.land/anthropics-ceo-drew-two-red-lines-the-pentagon-said-no/
• TechPolicy.Press, “A Timeline of the Anthropic-Pentagon Dispute”: https://www.techpolicy.press/a-timeline-of-the-anthropic-pentagon-dispute/
• The Conversation, “From Anthropic to Iran: Who Sets the Limits on AI’s Use in War and Surveillance?”: https://theconversation.com/from-anthropic-to-iran-who-sets-the-limits-on-ais-use-in-war-and-surveillance-277334

OpenAI Military Contracts

• The Intercept, “OpenAI on Surveillance and Autonomous Killings: You’re Going to Have to Trust Us”: https://theintercept.com/2026/03/08/openai-anthropic-military-contract-ethics-surveillance/
• Jacobin, “OpenAI Is Bleeding Cash. Its Solution? Military Contracts.”: https://jacobin.com/2026/04/openai-defense-contracts-tech-militarism
• Lawfare, “Military AI Policy by Contract: The Limits of Procurement as Governance”: https://www.lawfaremedia.org/article/military-ai-policy-by-contract–the-limits-of-procurement-as-governance
• Seeking Alpha, “OpenAI Secures $200 Million Pentagon Contract”: https://seekingalpha.com/news/4458694-openai-secures-200-million-pentagon-contract-to-deliver-ai-solutions-to-us-defense-department

Palantir Surveillance and Weapons

• SEC Filing, Palantir PX14A6G (human rights risk documentation): https://www.sec.gov/Archives/edgar/data/0001321655/000121465926005220/o429261px14a6g.htm
• Press TV, “Algorithm of War: How Palantir Became Pentagon’s Indispensable AI Arsenal”: https://www.presstv.ir/Detail/2026/04/26/767567/algorithm-war-how-palantir-became-pentagon-indispensable-ai-arsenal-wars-abroad

Palantir-Google Partnership

• Investing.com, “Palantir Partners with Google Cloud on Data Integration”: https://www.investing.com/news/company-news/palantir-partners-with-google-cloud-on-data-integration-93CH-4726337
• Yahoo Finance, “Palantir Unveils Google Cloud Partnership at AIPCon”: https://finance.yahoo.com/markets/stocks/articles/palantir-unveils-google-cloud-partnership-171316169.html
• Rolling Out, “Palantir and Google Cloud Just Made a Powerful AI Move”: https://rollingout.com/2026/06/04/palantir-and-google-cloud-just-made-ai/
• Yahoo Finance, “Palantir Partners with Google Cloud to Integrate Gemini AI Tools”: https://finance.yahoo.com/markets/stocks/articles/palantir-partners-google-cloud-integrate-193307429.html
• Crypto Briefing, “Palantir Partners with Google Cloud to Enhance Platform Integrations”: https://cryptobriefing.com/palantir-google-cloud-partnership-integrations/

Apple Siri AI / Gemini

• Tech Insider, “WWDC 2026: Siri AI Runs on Google’s $1B Gemini Deal”: https://tech-insider.org/wwdc-2026-siri-ai-gemini-deal/
• MacRumors, “Google Confirms Gemini-Powered Siri Coming Later This Year”: https://www.macrumors.com/2026/04/22/google-gemini-powered-siri-2026/
• 9to5Mac, “Apple Unveils New Siri AI in iOS 27”: https://9to5mac.com/2026/06/08/new-siri-whats-new/
• Apple Newsroom, “Apple Introduces Siri AI”: https://www.apple.com/newsroom/2026/06/apple-introduces-siri-ai-a-profoundly-more-capable-and-personal-assistant/
• MacRumors, “Apple Announces New CarPlay Features on iOS 27”: https://www.macrumors.com/2026/06/08/new-apple-carplay-features-ios-27/
• Auto Express, “Huge Apple CarPlay Update Revealed: Siri AI”: https://www.autoexpress.co.uk/news/369757/huge-apple-carplay-update-revealed-siri-ai-improved-maps-audio-player-and-more

Photo by LOGAN WEAVER | @LGNWVR on Unsplash