The cleaner the separation between the automatable task and the human judgment layer, the safer - and often more valuable - the profession becomes. A field like translation demonstrates what happens when that distinction collapses - not a sudden shift, but a quiet repricing.
Emerging companies like Palabra.ai are doing in real time what once required rooms full of human interpreters. Meanwhile, Harlequin France - publisher of titles like Médecins et Célibataires and Passion Pour un Inconnu - recently confirmed it is ending contracts with human translators entirely in favor of machine translation, leaving "freelance proofreaders" to sculpt the results. Those who remain aren't doing less work - rates per minute of video have collapsed from $5 to $1.50, while translators report the post-editing work is two to three times more cognitively demanding than translating from scratch. As Sarah O'Connor reported in the Financial Times, experienced translators are refusing MTPE work entirely, and as they exit, less skilled workers are stepping in.
The reason is structural. AI removed the creative problem-solving - thinking deeply about how to convey meaning, nuance, and cultural context - and left the routine. The automatable task was too close to the core deliverable.
Cybersecurity is the opposite story.
Here, the same technology arming the defense is simultaneously arming the attack. Anyone with access to an AI tool can now generate convincing phishing messages, automate malware, and run personalized deepfake attacks at a scale that previously required sophisticated criminal organizations. The barrier to entry for cybercrime has collapsed in almost exactly the way the barrier to entry for AI-assisted work has collapsed everywhere else - except here, the stakes are existential for the organizations caught in the middle.
The reason demand holds is mechanical. Every new AI deployment introduces attack surfaces that didn't exist before - model theft, prompt injection, agent hijacking. The organizations deploying AI fastest are, predictably, the most exposed.
White House chief of staff meets with Anthropic CEO over its new AI technology
The result is a threat landscape expanding faster than any automated defense can absorb. David Cass, President of CISOs Connect, puts it plainly: companies can lose north of $25 million in under 30 minutes. That's not a problem you solve with better algorithms alone. It requires human judgment, organizational context, and accountability that no AI can own.
This is where the profession's entry-level story gets interesting - and complicated. The traditional Tier-1 SOC analyst role, built around alert triage, log correlation, and incident documentation, is being automated out of existence. Many organizations have already merged Tier-1 and Tier-2 responsibilities entirely. What replaces it isn't fewer jobs - the field projects 29% growth through 2034. What replaces it is a fundamentally different kind of entry-level work.
The role is shifting to be one of partnering and advising, says Richard Watson, global cybersecurity consulting leader at EY, because a lot of the technology is doing the monitoring, triaging, quarantining, and so on.
Junior analysts stop being human filters. They become institutional translators - understanding why a specific threat matters inside this organization, with this configuration, and this risk appetite.
There is one honest caveat. The Tier-1 role wasn't just an entry point - it was the profession's training ground. The repetitive work of sifting through alerts and building pattern recognition over thousands of incidents is how junior analysts became senior ones. Automating that pipeline raises a question the industry hasn't fully answered: how does the next generation of leaders get built without the rungs that historically produced them?
The market isn't waiting. Take RunSybil, Terra Security, Xbow, and Novee - a new category of AI-native security startups that has collectively raised hundreds of millions in the past eighteen months, all built on the same thesis: automate the penetration testing, free the human analyst for the work that actually requires judgment.
AI handles the known, the routine, and the high-volume. What it cannot handle is the novel, the human-targeted, and the organizationally-specific - and the attack surface in all three categories is expanding faster than any wall can cover, because the same AI making defense cheaper is making sophisticated, targeted attacks cheaper to launch too.