Dunia Tidak Bersedia untuk Kecemasan AI

(SeaPRwire) –   Bayangkan bangun tidur dan mendapati internet berkelip-kelip, pembayaran kad gagal, ambulans pergi ke alamat yang salah, dan siaran kecemasan yang anda tidak lagi pasti boleh dipercayai. Sama ada disebabkan oleh kerosakan model, penggunaan jenayah, atau kejutan siber yang semakin meningkat, krisis yang didorong oleh AI boleh bergerak merentasi sempadan dengan cepat.

Dalam kebanyakan kes, tanda-tanda pertama kecemasan AI mungkin kelihatan seperti gangguan umum atau kegagalan keselamatan. Hanya kemudian, jika ada, barulah jelas bahawa sistem AI telah memainkan peranan penting.

Sesetengah kerajaan dan syarikat telah mula membina pagar kawalan untuk menguruskan kecemasan seperti ini. The European Union AI Act, the United States National Institute of Standards and Technology risk framework, the G7 Hiroshima process and international technical standards all aim to prevent harm. Cybersecurity agencies and infrastructure operators also have runbooks for hacking attempts, outages, and routine system failures. What is missing is not the technical playbook for patching servers or restoring networks. It is the plan for preventing social panic and a breakdown in trust, diplomacy, and basic communication if AI sits at the center of a fast-moving crisis. 

Mencegah kecemasan AI hanya separuh tugas. Separuh yang hilang dalam tadbir urus AI adalah kesiapsiagaan dan tindak balas. Siapa yang memutuskan bahawa insiden AI telah menjadi kecemasan antarabangsa? Siapa yang berbicara dengan orang ramai apabila mesej palsu membanjiri suapan mereka? Siapa yang mengekalkan saluran terbuka antara kerajaan jika talian normal terganggu?

Kerajaan boleh, dan mesti, menetapkan pelan tindak balas kecemasan AI sebelum terlambat. Dalam penyelidikan akan datang berdasarkan undang-undang bencana dan pengalaman dari kecemasan global lain, saya mengkaji bagaimana peraturan antarabangsa sedia ada sudah mengandungi komponen untuk playbook AI. Kerajaan sudah mempunyai alat undang-undang, tetapi kini perlu mencapai persetujuan bagaimana dan bila menggunakan mereka. Kita tidak memerlukan institusi baru yang rumit untuk mengawasi AI—kita hanya perlu kerajaan merancang terlebih dahulu. 

Cara Menyiapkan Diri untuk Kecemasan AI

Kita telah melihat model tadbir urus umum sebelum ini. The International Health Regulations membenarkan World Health Organisation mengisytiharkan kecemasan kesihatan global dan menyelaraskan tindakan. Nuclear accident treaties require rapid notification when radiation could spread across borders. Telecommunications agreements clear legal barriers so emergency satellite equipment can be switched on quickly. Cybercrime conventions set up 24/7 contact points so police forces can co-operate at short notice. The lessons show pre-agreed triggers, named co-ordinators, and fast communication channels save time in an emergency.

Kecemasan AI memerlukan asas yang sama. Mulakan dengan definisi yang dikongsi. An AI emergency should be an extraordinary event caused by the development, use, or malfunction of AI that risks severe cross-border harm and outstrips any single country’s capacity to cope. Crucially it must also cover situations where AI involvement is only suspected or is one of several plausible causes so that governments can act before forensic certainty arrives, if it arrives at all. Most incidents will never reach that level. Agreeing the definition in advance helps avoid paralysis during the first critical hours.

Seterusnya, kerajaan memerlukan playbook yang praktikal. The first element of this playbook should be defining a common set of triggers and a basic severity scale so officials know when to escalate from routine incident to international alert, including criteria for determining where AI involvement is only credibly suspected rather than conclusively proven. A second chapter should include naming a global co-ordinator who can convene quickly, supported by technical experts, law enforcement partners and disaster specialists. A third part should be establishing interoperable incident reporting systems so countries and companies can exchange essential information in minutes, not days. Next, we must create crisis communication protocols using authenticated, analogue methods such as radio. Finally, we must write a clear list of continuity and containment measures. These might include slowing high-risk AI services or switching critical infrastructure to manual control.

Menyusun Kesiapsiagaan Kecemasan AI

Jadi, siapa yang harus mengawasi inisiatif kesiapsiagaan kecemasan AI ini? Jawapan saya: the United Nations. 

Menempatkan sistem ini dalam struktur PBB penting untuk beberapa alasan. One is that an AI emergency will not respect alliances. A UN anchored mechanism offers wider inclusion and reduces duplication among rival coalitions. It provides technical help to countries without advanced AI capacity so the burden is not carried by a handful of major powers. It adds legitimacy and constraint. Extraordinary powers must be lawful, proportionate, and reviewable, especially when they touch digital networks used by billions of people.

Lapisan antarabangsa ini perlu dipadankan dengan langkah domestik yang boleh diambil oleh kerajaan sekarang. Every country should name a 24/7 AI emergency contact point. Emergency powers should be reviewed to see whether they cover AI infrastructure. Sector plans should be aligned with basic incident management and business continuity standards. Joint exercises should practice disinformation waves, model failures, and cross sector outages. Migration to post-quantum cryptography should be prioritised before a hostile attack forces such an update. Governments should also register trusted senders and alert templates so messages can still reach citizens when systems are unstable.

Langkah-langkah pencegahan ini adalah perlu . Serangan siber yang berkaitan AI yang dilaporkan adalah dan banyak negara sudah mengalami gangguan skala kecil, percubaan manipulasi data, dan lonjakan disinformasi yang memberi petunjuk tentang bagaimana peristiwa yang lebih besar boleh kelihatan. What’s more, a fast-moving AI failure could combine with today’s hyper-connected infrastructure to produce a crisis that no single country can handle alone.

Ini bukan panggilan untuk agensi super global baru. It is a call to stitch together what already exists into a coherent response. We need an AI emergency playbook that borrows these tools and rehearses them.

Ukuran tadbir urus AI akan menjadi bagaimana kita bertindak balas pada hari terburuk kita. Pada masa kini, dunia tidak mempunyai pelan untuk kecemasan AI—tetapi kita boleh menciptanya. Kita mesti membina ia sekarang, mengujinya, dan mengikatnya kepada undang-undang dengan langkah pengawasan, kerana sekali krisis seterusnya telah bermula, ia sudah terlambat.