Jakarta, CMNexus. Adopsi kecerdasan artifisial (AI) di perusahaan-perusahaan Asia Pasifik, termasuk Indonesia, terus melaju. Namun laporan terbaru dari pusat penelitian internal Kaspersky menunjukkan akselerasi itu membawa risiko keamanan siber baru: penggunaan alat AI pihak ketiga oleh karyawan tanpa pengawasan resmi, atau yang dikenal sebagai shadow AI.
Temuan ini penting bagi pelaku usaha, dari korporasi sampai usaha kecil yang mulai memakai alat AI untuk membuat konten, merangkum dokumen, hingga melayani pelanggan. Semakin banyak pekerjaan yang dipindahkan ke alat AI, semakin besar pula data yang berpindah keluar dari kendali perusahaan.
Kerentanan AI mulai diperhitungkan serius
Riset Kaspersky yang dipublikasikan pada 9 September 2026 mencatat 13% organisasi di Asia Pasifik menempatkan kerentanan AI sebagai salah satu ancaman utama bagi sistem keamanan siber mereka. Angka itu masih di bawah eksploitasi celah perangkat lunak (20%) dan serangan phishing (19%), tetapi posisinya kini naik menjadi pertimbangan kritis dalam operasional bisnis modern di kawasan.
Tingkat kekhawatiran berbeda antarnegara: India mencatat proporsi tertinggi (20%), disusul Malaysia (15%), Thailand dan Vietnam (masing-masing 13%), kemudian Indonesia (10%), dan Tiongkok (7%).
Group Manager Kaspersky AI Technology Research Centre, Vladislav Tushkanov, menyatakan percepatan pemanfaatan AI untuk memangkas durasi proses bisnis berdampak langsung pada meningkatnya potensi serangan siber.
"Seiring organisasi secara aktif mengadopsi AI untuk mempercepat proses bisnis dan mengurangi beban kerja rutin, risiko keamanan seperti kerentanan AI dan shadow AI juga meningkat," ujarnya dalam laporan riset resmi tersebut.
Shadow AI: 24% organisasi di Indonesia melaporkan
Shadow AI menggambarkan kebiasaan karyawan memakai alat AI pihak ketiga — dari ChatGPT hingga DeepSeek — untuk tugas kerja tanpa izin resmi departemen TI. Praktik ini membuka jalan kebocoran data sensitif, mulai dari dokumen internal, data pelanggan, sampai informasi komersial.
Hasil survei menunjukkan sekitar 30% responden di Asia Pasifik menilai tren shadow AI sebagai faktor risiko yang sangat umum. Sebarannya: India (36%), Malaysia (35%), Vietnam (33%), Tiongkok (31%), Indonesia (24%), dan Thailand (18%).
Bagi banyak usaha kecil dan menengah yang belum punya kebijakan tertulis soal AI, angka itu menjadi pengingat bahwa alat AI yang dipakai tim sehari-hari sering kali berada di luar pantauan pemilik usaha.
60% organisasi memilih mengizinkan dengan pembatasan
Alih-alih melarang total, mayoritas perusahaan di kawasan memilih pendekatan moderat. Sebanyak 60% organisasi di Asia Pasifik mengizinkan penggunaan platform AI pihak ketiga dengan pembatasan ketat — misalnya melarang mengunggah nama perusahaan atau dokumen internal — serta menyediakan pedoman dan pelatihan penggunaan yang aman. Kebijakan serupa tercatat di Tiongkok (62%), serta India, Malaysia, dan Vietnam (masing-masing 61%).
Adopsi AI internal tetap melaju
Di sisi lain, integrasi AI ke dalam organisasi tidak melambat. Sebanyak 85% perusahaan di Asia Pasifik sedang mendiskusikan, mengembangkan, menerapkan, atau menguji coba alat AI berbasis Large Language Model (LLM), sementara 10% lainnya sudah mengintegrasikan teknologi tersebut ke dalam alur kerja.
Pendorong utamanya adalah peningkatan kualitas pengambilan keputusan, efisiensi proses, dan pengurangan kesalahan manusia. Secara global, survei yang sama menemukan 41% perusahaan menyatakan adopsi AI sebagai pendorong utama peningkatan investasi keamanan informasi mereka — bersaing dengan kenaikan biaya akibat insiden kebocoran data, migrasi ke infrastruktur cloud, dan insiden keamanan yang baru saja dialami.
Regulasi AI Indonesia di depan mata
Temuan ini muncul bersamaan dengan penyiapan kerangka regulasi AI nasional. Kementerian Komunikasi dan Digital menyebut dua Rancangan Peraturan Presiden — tentang etika kecerdasan artifisial dan peta jalan kecerdasan artifisial nasional — sudah memasuki tahap akhir dan tinggal menunggu tanda tangan Presiden, dengan target terbit tahun ini.
Direktur Jenderal Ekosistem Digital Komdigi, Edwin Hidayat Abdullah, menyatakan industri menantikan aturan tersebut sebagai panduan arah pengembangan AI. Setelah Perpres berlaku, kementerian dan lembaga akan diminta mengatur pemanfaatan AI sesuai bidang masing-masing, termasuk penerapan pendekatan berbasis tingkat risiko penggunaan.
Langkah praktis untuk pelaku usaha
Beberapa langkah berikut bisa menjadi titik awal tata kelola AI di perusahaan, besar maupun kecil:
- Inventarisasi alat AI yang benar-benar dipakai tim, termasuk akun pribadi yang dibawa masuk ke pekerjaan.
- Susun kebijakan tertulis yang jelas: data apa yang boleh dan tidak boleh diunggah ke alat AI publik.
- Klasifikasi data — pisahkan data publik, internal, dan sensitif (data pelanggan, keuangan, kontrak).
- Latih karyawan agar memahami risiko dan cara aman memakai alat AI.
- Kurangi alat yang tidak terkendali: pilih layanan dengan pengaturan kontrol data dan catat penggunaannya.
Kecepatan adopsi AI tidak harus dikorbankan. Yang dibutuhkan adalah kontrol yang sepadan, agar produktivitas naik tanpa membuka celah kebocoran data baru.
Ikuti perkembangan teknologi dan bisnis lainnya di kanal berita CMNexus.
Sumber: - Kaspersky, siaran pers "AI adoption is the main driver for companies boosting their IT Security budgets" (29 Juli 2026): https://www.kaspersky.com/about/press-releases/ai-adoption-is-the-main-driver-for-companies-boosting-their-it-security-budgets - Bisnis.com, "Kaspersky Ungkap Risiko Shadow AI Mengintai Korporasi Asia Pasifik" (9 September 2026): https://teknologi.bisnis.com/read/20260909/84/2002896/kaspersky-ungkap-risiko-shadow-ai-mengintai-korporasi-asia-pasifik - ITBeat, "Kaspersky Highlights Shadow AI Amid Rising Security Investments" (9 September 2026): https://www.itbeat.id/en/kaspersky-soroti-shadow-ai-di-tengah-peningkatan-investasi-keamanan/ - Pasardana, "Adopsi AI Melesat, Ancaman Shadow AI Mengintai Perusahaan Asia Pasifik" (9 September 2026): https://pasardana.id/news/2026/9/9/adopsi-ai-melesat-ancaman-shadow-ai-mengintai-perusahaan-asia-pasifik - Indotelko, "Shadow AI jadi risiko baru perusahaan di Asia Pasifik" (9 September 2026): https://www.indotelko.com/read/1788962137/shadow-ai-jadi-risiko-baru-perusahaan-di-asia-pasifik - Kompas Tekno, "Komdigi: Perpres AI Tinggal Tunggu Tanda Tangan Presiden, Target Terbit Tahun Ini" (3 September 2026): https://tekno.kompas.com/read/2026/09/03/20290047/komdigi--perpres-ai-tinggal-tunggu-tanda-tangan-presiden-target-terbit-tahun - ANTARA, "Perpres AI dinantikan industri jadi arah pengembangan AI nasional": https://www.antaranews.com/berita/5724532/perpres-ai-dinantikan-industri-jadi-arah-pengembangan-ai-nasional
Jakarta, CMNexus. The adoption of artificial intelligence (AI) across Asia-Pacific companies, including in Indonesia, continues to accelerate. However, a recent report from Kaspersky's internal research center shows that this rapid expansion brings a new cybersecurity risk: the use of third-party AI tools by employees without official oversight, commonly known as shadow AI.
These findings are critical for business leaders, from large corporations to small enterprises that have begun utilizing AI tools to generate content, summarize documents, and handle customer service. As more tasks are offloaded to AI tools, a greater volume of data moves outside company control.
AI vulnerabilities are now taken seriously
Kaspersky research published on September 9, 2026, notes that 13% of organizations in the Asia-Pacific rank AI vulnerabilities among the primary threats to their cybersecurity systems. While this figure still trails software vulnerability exploitation (20%) and phishing attacks (19%), it has elevated into a critical consideration in modern regional business operations.
Levels of concern vary by country: India recorded the highest proportion (20%), followed by Malaysia (15%), Thailand and Vietnam (13% each), then Indonesia (10%), and China (7%).
Vladislav Tushkanov, Group Manager at the Kaspersky AI Technology Research Centre, stated that the accelerated use of AI to shorten business processes directly impacts the rising potential for cyberattacks.
"As organizations actively adopt AI to speed up business processes and reduce routine workloads, security risks such as AI vulnerabilities and shadow AI are also on the rise," he said in the official research report.
Shadow AI: 24% of organizations in Indonesia report usage
Shadow AI describes the practice of employees using third-party AI tools—from ChatGPT to DeepSeek—for work-related tasks without formal authorization from the IT department. This practice paves the way for leaks of sensitive data, ranging from internal documents and customer data to commercial information.
Survey results indicate that around 30% of respondents in the Asia-Pacific consider the shadow AI trend to be a very common risk factor. The breakdown shows: India (36%), Malaysia (35%), Vietnam (33%), China (31%), Indonesia (24%), and Thailand (18%).
For many micro, small, and medium enterprises that lack written AI policies, this figure serves as a reminder that the AI tools their teams use daily often operate outside the business owner's radar.
60% of organizations choose permission with restrictions
Rather than imposing a blanket ban, the majority of companies in the region are adopting a moderate approach. As many as 60% of organizations in the Asia-Pacific permit the use of third-party AI platforms under strict restrictions—such as prohibiting the upload of company names or internal documents—while providing guidelines and safe-use training. Similar policies were recorded in China (62%), as well as India, Malaysia, and Vietnam (61% each).
Internal AI adoption continues to surge
On the other hand, the integration of AI into organizations is not slowing down. A total of 85% of companies in the Asia-Pacific are discussing, developing, deploying, or testing Large Language Model (LLM)-based AI tools, while another 10% have already integrated the technology into their workflows.
The primary drivers include enhanced decision-making quality, process efficiency, and reduced human error. Globally, the same survey found that 41% of companies cited AI adoption as the main driver for increasing their information security investments—competing with rising costs from data breach incidents, migration to cloud infrastructure, and recent security incidents.
Indonesian AI regulation on the horizon
These findings emerge alongside the preparation of Indonesia's national AI regulatory framework. The Ministry of Communication and Digital affairs (Komdigi) stated that two Draft Presidential Regulations—governing artificial intelligence ethics and the national artificial intelligence roadmap—have entered their final stages and are only awaiting the President's signature, targeting issuance this year.
Director General of Digital Ecosystem at Komdigi, Edwin Hidayat Abdullah, stated that the industry is eagerly anticipating these regulations as a guiding direction for AI development. Once the Presidential Regulation takes effect, ministries and agencies will be instructed to regulate AI utilization within their respective sectors, including applying a risk-level-based approach to its use.
Practical steps for business operators
Several practical steps can serve as a starting point for AI governance in companies, both large and small:
- Inventory AI tools actually being used by the team, including personal accounts brought into work.
- Draft a clear written policy: outline what data can and cannot be uploaded to public AI tools.
- Classify data—separate public, internal, and sensitive data (customer records, financials, contracts).
- Train employees to understand the risks and how to use AI tools safely.
- Reduce uncontrolled tools: choose services equipped with data control settings and log their usage.
The pace of AI adoption does not have to be sacrificed. What is needed is commensurate control, ensuring productivity rises without opening new avenues for data leaks.
Follow other technology and business developments on the CMNexus news channel.
Sources: - Kaspersky, press release "AI adoption is the main driver for companies boosting their IT Security budgets" (July 29, 2026): https://www.kaspersky.com/about/press-releases/ai-adoption-is-the-main-driver-for-companies-boosting-their-it-security-budgets - Bisnis.com, "Kaspersky Ungkap Risiko Shadow AI Mengintai Korporasi Asia Pasifik" (September 9, 2026): https://teknologi.bisnis.com/read/20260909/84/2002896/kaspersky-ungkap-risiko-shadow-ai-mengintai-korporasi-asia-pasifik - ITBeat, "Kaspersky Highlights Shadow AI Amid Rising Security Investments" (September 9, 2026): https://www.itbeat.id/en/kaspersky-soroti-shadow-ai-di-tengah-peningkatan-investasi-keamanan/ - Pasardana, "Adopsi AI Melesat, Ancaman Shadow AI Mengintai Perusahaan Asia Pasifik" (September 9, 2026): https://pasardana.id/news/2026/9/9/adopsi-ai-melesat-ancaman-shadow-ai-mengintai-perusahaan-asia-pasifik - Indotelko, "Shadow AI jadi risiko baru perusahaan di Asia Pasifik" (September 9, 2026): https://www.indotelko.com/read/1788962137/shadow-ai-jadi-risiko-baru-perusahaan-di-asia-pasifik - Kompas Tekno, "Komdigi: Perpres AI Tinggal Tunggu Tanda Tangan Presiden, Target Terbit Tahun Ini" (September 3, 2026): https://tekno.kompas.com/read/2026/09/03/20290047/komdigi--perpres-ai-tinggal-tunggu-tanda-tangan-presiden-target-terbit-tahun - ANTARA, "Perpres AI dinantikan industri jadi arah pengembangan AI nasional": https://www.antaranews.com/berita/5724532/perpres-ai-dinantikan-industri-jadi-arah-pengembangan-ai-nasional
