š¤ Hackers are starting to āteam up with AIā on the job: 400+ Latin American banks and payment apps targeted
According to The Information, cybersecurity firm Straiker revealed a hacking group called CyberCartel. From May to September this year, they used models like Claude, Grok, and Kimi to build an AI agent system that divided up tasks and coordinated automated attacks on more than 400 financial targets in Latin America, including banks and payment apps.
A few details worth noting:
š¹ They disguised their requests as āsecurity testingā to get around the modelsā safety restrictions
š¹ Their main loot was passwords, API keys, and customer dataānot some god-tier vulnerability
š¹ Anthropic has blocked the related access, while xAI says it is investigating
š¹ So far, most of the information comes from Straiker alone, and the details still need further verification
My take: What should make ordinary people wary isnāt āhow smart AI is,ā but that the cost of launching attacks has gone down. Hackers used to have to pick the big fish. Now AI can try and crack accounts in bulk, making even small accounts worth a quick attempt.
In crypto, the most common points of entry are still the same old problems:
Using the same password for exchanges, email, and social accounts;
Leaving API keys lying around in the cloud or chat logs;
Only enabling SMS for two-factor authentication.
Do a quick security check: use a unique password for each exchange and a password manager; prioritize passkeys or an authenticator app; enable only the necessary API permissions and restrict access by IP; and secure your email account even more than your exchange account, because it can reset everything.
How many platforms are you still using the same password for? Be honestāthe comments wonāt judge you š
#AIå®å Ø #蓦ę·å®å Ø #å åÆ蓧åø #BinanceSquare
According to The Information, cybersecurity firm Straiker revealed a hacking group called CyberCartel. From May to September this year, they used models like Claude, Grok, and Kimi to build an AI agent system that divided up tasks and coordinated automated attacks on more than 400 financial targets in Latin America, including banks and payment apps.
A few details worth noting:
š¹ They disguised their requests as āsecurity testingā to get around the modelsā safety restrictions
š¹ Their main loot was passwords, API keys, and customer dataānot some god-tier vulnerability
š¹ Anthropic has blocked the related access, while xAI says it is investigating
š¹ So far, most of the information comes from Straiker alone, and the details still need further verification
My take: What should make ordinary people wary isnāt āhow smart AI is,ā but that the cost of launching attacks has gone down. Hackers used to have to pick the big fish. Now AI can try and crack accounts in bulk, making even small accounts worth a quick attempt.
In crypto, the most common points of entry are still the same old problems:
Using the same password for exchanges, email, and social accounts;
Leaving API keys lying around in the cloud or chat logs;
Only enabling SMS for two-factor authentication.
Do a quick security check: use a unique password for each exchange and a password manager; prioritize passkeys or an authenticator app; enable only the necessary API permissions and restrict access by IP; and secure your email account even more than your exchange account, because it can reset everything.
How many platforms are you still using the same password for? Be honestāthe comments wonāt judge you š
#AIå®å Ø #蓦ę·å®å Ø #å åÆ蓧åø #BinanceSquare