Tech giants take malicious AI seriously: what smaller businesses should borrow from their approach

ER10 reported that leading technology companies have significantly stepped up work against malicious AI — from phishing written by neural networks to automated vulnerability hunting. The reason is prosaic: tools that speed up honest work speed up attacks just as well. A scammer's email is now literate, personalized and arrives in thousands of variants.

What matters is how the giants respond: not with a single «buy protection» move but by changing assumptions. The baseline scenario shifts from «we are unlikely to be attacked» to «attacks are constant, cheap and convincing» — and processes are rebuilt around it: verification of payment and bank-detail-change requests, access restrictions, incident response rehearsed in advance.

For a company in Kazakhstan this is precisely the borrowable part. You will not have a tech giant's budget, but their assumptions are free: treat a convincing «bank» email as the norm, not a rarity; require a second confirmation channel for money and access; know in advance who does what during an incident. AI made attacks mass-scale — but the same processes that catch human errors, properly tuned, catch machine ones too.