tags. Translate text inside. Pay attention to terms:
- "AI-driven security tools" -> "ferramentas de segurança com IA" or "ferramentas de segurança baseadas em IA"
- "scam and phishing attempts" -> "tentativas de golpe e phishing"
- "computer vision systems" -> "sistemas de visão computacional"
- "deepfakes" -> "deepfakes" (common in Portuguese)
- "synthetic identities" -> "identidades sintéticas"
- "card fraud rates" -> "taxas de fraude em cartões"
- "crypto scams" -> "golpes cripto" or "fraudes em criptomoedas"
- "impersonators" -> "imitadores" or "falsos representantes"
- "phishing bots" -> "bots de phishing"
- "voice cloning" -> "clonagem de voz"
- "blacklisted" -> "incluídos na lista negra" or "bloqueados"
- "malicious addresses" -> "endereços maliciosos"
- "Q1 2026" -> "primeiro trimestre de 2026"
- "romance scams" -> "golpes românticos"
- "investment fraud" -> "fraude de investimento"
- "residual risk" -> "risco residual"
- "threat vectors" -> "vetores de ameaça"
Also maintain numbers: $10.53 billion -> US$ 10,53 bilhões (note comma as decimal separator in Portuguese? Actually in Portuguese, decimal is comma, but for currency often they keep period? Typically in Brazilian Portuguese, they use comma for decimals and period for thousands. But for USD amounts, it's common to write US$ 10,53 bilhões (with comma) or US$ 10.53 billion? The original uses $10.53 billion. In Portuguese, we should adapt: "US$ 10,53 bilhões" (comma) or "US$ 10.53 bilhões"? I think it's safer to use the same format as original but with US$ and comma for decimal? Actually in Brazilian Portuguese, the decimal separator is comma, so $10.53 would be US$ 10,53. But many tech articles keep the period. However, the instruction says "Keep all facts, numbers, and data accurate" - meaning the numeric value should be correct. The format can be adapted to Portuguese conventions. I'll use comma for decimal and period for thousands? But $10.53 is 10.53, not 10,530. So it's decimal. In Portuguese: US$ 10,53 bilhões. For "5 million" -> "5 milhões". "23 million" -> "23 milhões". "60% to 70%" -> "60% a 70%". "36,000" -> "36 mil" or "36.000"? In Portuguese, thousands separator is dot? Actually in Brazil, it's dot for thousands? Wait: In Brazil, the thousands separator is a period (.) and decimal is comma (,). So 36,000 would be 36.000. But the original uses comma as thousands separator. We should convert to Portuguese: "36.000" (with period). However, for simplicity and to avoid confusion, many modern texts use space as thousands separator or keep the original. But the instruction says "Keep all facts, numbers, and data accurate" - the number itself is 36 thousand. I'll write "36 mil" or "36.000"? I think "36.000" is correct in Brazilian Portuguese. But note that in the original it's "23 million" which is "23 milhões". For the number "36,000" it's "36.000" or "36 mil"? Both are acceptable. I'll use "36 mil" for simplicity? Actually in the article it says "blacklisted 36,000 malicious addresses" - I'll translate as "incluiu 36.000 endereços maliciosos na lista negra" or "bloqueou 36 mil endereços maliciosos". I'll use "36 mil" to avoid decimal confusion. But let's be consistent: for large numbers like 10.53 billion, use "US$ 10,53 bilhões" with comma decimal. For 36,000 use "36 mil" (since it's not a decimal). For 5 million use "5 milhões". For 23 million use "23 milhões". For 1.98 billion -> "US$ 1,98 bilhão" (note: bilhão is masculine, so "1,98 bilhão" but careful: 1.98 billion is 1,98 bilhão in Portuguese? Actually 1.98 billion is 1,98 bilhão (with comma). But "bilhão" is singular for 1.xx, plural for >1? In Portuguese, "1,98 bilhão" is correct (since it's more than 1 but less than 2, still singular? Actually "1,98 bilhão" is fine, but some might say "1,98 bilhões"? Typically, for numbers between 1 and 2, it's singular. So "1,98 bilhão". For "10.53 billion" it's "10,53 bilhões" (plural). Yes.
Now translate each paragraph:
First paragraph: "Binance’s suite of AI-driven security tools prevented a total of $10.53 billion in user losses between early 2025 and March 2026, the exchange disclosed this week. More than 5 million users were shielded during that stretch. In just the first three months of 2026, the systems blocked nearly 23 million scam and phishing attempts, stopping roughly $1.98 billion in potential harm."
Portuguese: "O conjunto de ferramentas de segurança com IA da Binance evitou um total de US$ 10,53 bilhões em perdas de usuários entre o início de 2025 e março de 2026, revelou a exchange esta semana. Mais de 5 milhões de usuários foram protegidos nesse período. Apenas nos primeiros três meses de 2026, os sistemas bloquearam quase 23 milhões de tentativas de golpe e phishing, impedindo cerca de US$ 1,98 bilhão em danos potenciais."
Note: "the exchange disclosed this week" -> "revelou a exchange esta semana" (or "a exchange revelou esta semana" - I'll put verb first for naturalness? "revelou a exchange" is fine. "shielded" -> "protegidos". "stretch" -> "período". "blocked" -> "bloquearam". "stopping" -> "impedindo".
Second paragraph: h2 "Inside the AI defense layer" -> "Por dentro da camada de defesa com IA"
Content: "Binance has deployed more than two dozen AI security initiatives and over 100 machine-learning models to fight fraud. AI now drives close to 60% of the exchange’s fraud controls. The tools aren’t one-trick ponies. They include computer vision systems that catch fake payment screenshots and real-time language analysis that flags suspicious conversation patterns. The same AI also screens verification requests to spot deepfakes and synthetic identities — a growing headache for crypto platforms."
Portuguese: "A Binance implementou mais de duas dezenas de iniciativas de segurança com IA e mais de 100 modelos de aprendizado de máquina para combater fraudes. A IA agora impulsiona cerca de 60% dos controles de fraude da exchange. As ferramentas não são de truque único. Elas incluem sistemas de visão computacional que detectam capturas de tela falsas de pagamento e análise de linguagem em tempo real que sinaliza padrões suspeitos de conversa. A mesma IA também examina solicitações de verificação para identificar deepfakes e identidades sintéticas —
