Source Details

{
    "source_type": "extension",
    "content_hash": "f5af44da228e8eacb034d571cbd7ad5ebe17639d0832a38cfd59fdaa72b359ff",
    "submitted_via": "chrome_extension",
    "extension_version": "1.0.18",
    "original_url": "https:\/\/www.spiegel.de\/international\/world\/the-uncertain-future-what-happens-if-we-lose-control-of-ai-a-abb932c3-a696-4506-8175-fd940aec3dc5?utm_source=cnn_Global+Briefing&utm_medium=email",
    "parsed_content": "Zur Merkliste hinzuf\u00fcgen\nX.com\nFacebook\nE-Mail\nLink kopieren\nWeitere Optionen zum Teilen\nX.com\nFacebook\nE-Mail\nMessenger\nWhatsApp\nLink kopieren\nSPIEGEL bei Google bevorzugen\nSie k\u00f6nnen den Artikel leider nicht mehr aufrufen. Der Link, der Ihnen geschickt wurde, ist entweder \u00e4lter als 30 Tage oder der Artikel wurde bereits 10 Mal ge\u00f6ffnet.\nPeople have imagined machines taking over the world many times before. Daniel Kokotajlo has named his version Plan D. As in \"Destruction.\" The scenario where everything goes wrong, he says. With artificial intelligence. With humanity.Kokotajlo used to do research at OpenAI. Now, his organization, the AI Futures Project, tries to predict how the technology will develop. He believes there are several possible endings \u2013 among them a very positive one, Plan A, and a very negative one, Plan D.\nArtificial intelligence, something that perhaps gets a bit lost in the current debate, is not necessarily some evil force descending upon humanity. Like any technology, it can do good. But it can just as easily do harm. What matters is whom it serves. And whether people can control it. Right now, though, it is unfortunately Kokotajlo's dystopian Plan D that is gaining in probability.\nDER SPIEGEL 39\/2026\nThe article you are reading originally appeared in German in issue 39\/2026 (September 18th, 2026) of DER SPIEGEL.\nSPIEGEL International\nThe AI industry is being rattled by glaring security incidents. AI agents have reportedly gone rogue on their own. The heads of the largest AI companies are warning of existential risks to humanity, of a looming collapse of the internet. They are calling for binding global rules and government assistance. US President Donald Trump, meanwhile, wants no such laws. He wants US technological leadership at any price.\nThis fall could go down in history. As the moment humanity lost control of the super-machine it created.\nDER SPIEGEL relies on support from its readers. To enjoy the rest of this story, please consider signing up for a subscription to The German View, our English-language channel on Substack. The article can be found here\u00a0, and our Substack channel can be found here . Alternatively, you can sign up for a SPIEGEL+ subscription below.\nCan the risk be contained through intelligent regulation? How can AI systems be built so that they always operate in humanity's interest? What should governments and companies do?Or is it already too late? Are the machines already too intelligent to be controlled by an inferior species?\nDiesen Artikel weiterlesen mit SPIEGEL+\nSie haben bereits ein Digital-Abo?\nZum Login\nSPIEGEL plus\nNur f\u00fcr Neukunden\n1 \u20ac f\u00fcr 4 Wochen\ndanach 6,49 \u20ac pro Woche\nFreier Zugriff auf alle S+-Artikel auf SPIEGEL.de und in der App\nW\u00f6chentlich die digitale Ausgabe des SPIEGEL inkl. E-Paper (PDF), Digital-Archiv und S+-Newsletter\nJederzeit k\u00fcndigen\nJetzt abonnieren\nSPIEGEL plus\n52 Wochen 31 % sparen\n4,49 \u20ac pro Woche f\u00fcr 52 Wochen\ndanach 6,49 \u20ac pro Woche\nFreier Zugriff auf alle S+-Artikel auf SPIEGEL.de und in der App\nW\u00f6chentlich die digitale Ausgabe des SPIEGEL inkl. E-Paper (PDF), Digital-Archiv und S+-Newsletter\n52 Wochen rabattierte Laufzeit\nJetzt abonnieren\nSie haben bereits ein Print-Abo?\nHier rabattiert Digital-Zugang bestellen\nJetzt Artikel freischalten:\nSie haben bereits ein Digital-Abo?\nZum Login\nSPIEGEL plus\nundefined pro Monat\nZugang zu allen Artikeln in der App und auf SPIEGEL.de\nW\u00f6chentliche Ausgabe des SPIEGEL als E-Paper\nJederzeit k\u00fcndbar\nJetzt abonnieren\nSPIEGEL plus\nundefined sparen\nundefined pro Jahr\nZugang zu allen Artikeln in der App und auf SPIEGEL.de\nW\u00f6chentliche Ausgabe des SPIEGEL als E-Paper\nJederzeit k\u00fcndbar\nJetzt abonnieren\niTunes-Abo wiederherstellen\nSPIEGEL+ wird \u00fcber Ihren iTunes-Account abgewickelt und mit Kaufbest\u00e4tigung bezahlt. 24 Stunden vor Ablauf verl\u00e4ngert sich das Abo automatisch um\neinen Monat zum Preis von zurzeit \nein Jahr zum Preis von zurzeit . In den Einstellungen Ihres iTunes-Accounts k\u00f6nnen Sie das Abo jederzeit k\u00fcndigen. Um SPIEGEL+ au\u00dferhalb\ndieser App zu nutzen, m\u00fcssen Sie das Abo direkt nach dem Kauf mit einem SPIEGEL-ID-Konto verkn\u00fcpfen. Mit dem Kauf akzeptieren Sie unsere\nAllgemeinen Gesch\u00e4ftsbedingungen und Datenschutzerkl\u00e4rung.\nStartseite\nFeedback",
    "published_at": "2026-09-17T16:08:00+02:00",
    "author_name": "Bernhard Zand, Marcel Rosenbach, Angela Gruber, Simon Book, DER SPIEGEL",
    "site_name": "SPIEGEL International",
    "section": null,
    "publisher": "DER SPIEGEL",
    "ai_headline": "The Uncertain Future: What Happens If We Lose Control of AI?",
    "ai_simplified_title": "Daniel Kokotajlo Warns of AI Destruction Scenario",
    "ai_excerpt": "Daniel Kokotajlo, formerly of OpenAI, warns of a 'Plan D' scenario where humanity loses control of AI. The article explores the potential for AI to do harm and the urgent need for regulation.",
    "ai_subject_tags": [
        "Artificial Intelligence",
        "AI Safety",
        "Technological Risk",
        "Future of AI",
        "Regulation"
    ],
    "ai_context_type": "Analysis",
    "ai_context_details": {
        "tone": "cautionary",
        "perspective": "critical",
        "audience": "general",
        "credibility_indicators": [
            "expert_quotes"
        ]
    },
    "ai_source_vector": null,
    "ai_confidence_score": 0.9999999999999999,
    "ai_extraction_metadata": {
        "extracted_at": "2026-09-30T12:04:59.973078Z",
        "ai_model": "gemini-2.5-flash",
        "extraction_method": "automated",
        "content_length": 4273,
        "metadata_prompt_body_chars": 4273,
        "metadata_prompt_truncated": false,
        "url": "https:\/\/spiegel.de\/international\/world\/the-uncertain-future-what-happens-if-we-lose-control-of-ai-a-abb932c3-a696-4506-8175-fd940aec3dc5",
        "existing_metadata": {
            "author_name": null,
            "published_at": null,
            "domain_name": "spiegel.de",
            "site_name": null,
            "section": null,
            "publisher": null
        },
        "openai_refusal_gemini_fallback": false
    }
}

{
    "extracted_at": "2026-09-30T12:04:59.973078Z",
    "ai_model": "gemini-2.5-flash",
    "extraction_method": "automated",
    "content_length": 4273,
    "metadata_prompt_body_chars": 4273,
    "metadata_prompt_truncated": false,
    "url": "https:\/\/spiegel.de\/international\/world\/the-uncertain-future-what-happens-if-we-lose-control-of-ai-a-abb932c3-a696-4506-8175-fd940aec3dc5",
    "existing_metadata": {
        "author_name": null,
        "published_at": null,
        "domain_name": "spiegel.de",
        "site_name": null,
        "section": null,
        "publisher": null
    },
    "openai_refusal_gemini_fallback": false
}