Even Programmers Aren’t Safe: The 10 Jobs AI Is Coming For First (Anthropic & Microsoft Report)

Summary

The video argues that AI exposure changes how jobs are performed; it does not automatically mean immediate job loss. The greatest near-term pressure is on routine, repeatable tasks and entry-level hiring. Workers who combine AI use with judgment, domain knowledge, communication, and problem-solving are better positioned than those whose value mainly comes from producing standard drafts, reports, or code.

Jobs discussed, from most to less exposed

1. Programmers: The presenter says an Anthropic report estimates about 75% task coverage for new software programmers. AI can generate code, but software engineering also involves solving unfamiliar problems. Recommended skills include debugging, databases, system design, architecture, and effective use of AI tools—not just memorizing syntax.
2. Financial analysts: AI can summarize reports, compare company performance, build models, and draft investment memos. Analysts still need accounting, valuation, risk analysis, and the ability to verify AI outputs.
3. Customer service and support: AI is already used for common requests such as password resets, order tracking, and FAQs. The video describes companies using retrieval-augmented generation (RAG) chatbots for basic support, while suggesting that more complex cases still require human judgment.
4. Sales and business development representatives: AI can identify prospects, find contact details, personalize outreach, and send emails. The human advantage lies in calls, negotiation, handling objections, building relationships, and closing enterprise deals.
5. Market research analysts: AI can summarize customer feedback and surveys, identify trends, and produce reports. A more durable skill is understanding customer behavior and interpreting what customers want—not simply compiling or presenting data.
6. Junior data analysts: Routine data organization, charting, and dashboard work are increasingly automatable. Analysts should learn SQL and data visualization, and develop the business understanding needed to explain why a metric changed and what action it suggests.
7. Technical writers: AI can draft documentation and create demos. The presenter suggests considering developer advocacy—including explaining technical products and building communities—as an alternative or complement to documentation-focused work.
8. Paralegals and legal researchers: AI can search documents and prepare first drafts, but may return outdated or jurisdictionally incorrect information. Human review and specialization in a legal domain remain important. The presenter mentions legal-AI companies including “Harv” (likely Harvey) and Legora.
9. Translators and interpreters: AI handles much routine translation but can miss emotional nuance and context. The recommendation is to specialize in an industry such as law or finance, where terminology and accuracy matter.
10. Public relations specialists: AI can quickly draft press releases, announcements, and campaign variations. PR still requires strategic judgment, media relationships, and careful decision-making during sensitive situations that could affect a brand’s reputation.

Frameworks, processes, and recommendations

• Separate task exposure from job replacement: Assess which parts of a role are repetitive and automatable, then focus on the judgment-heavy work that remains.
• Move from production to interpretation: Don’t stop at generating code, dashboards, translations, research summaries, or drafts. Add context, verification, and recommendations.
• Build domain expertise: Specialize in a field—such as law, finance, or a particular business function—to make AI-generated work more useful and reliable.
• Use AI while developing complementary skills: Learn relevant tools while strengthening problem-solving, communication, relationship-building, and business understanding.
• For data roles: Learn SQL and visualization, then connect findings to business causes and decisions.
• For sales roles: Use automation to handle prospecting and outreach efficiently; invest human effort in discovery, negotiation, objection handling, and closing.
• For technical communicators: Consider roles that combine product knowledge with explanation and community-building, such as developer advocacy.
• Verify consequential outputs: Especially in legal, financial, and other high-stakes work, check AI results for accuracy, currency, and context.

Metrics and claims cited in the video

• An Anthropic report is said to show about 75% task coverage for new software programmers.
• The presenter says automation workflows related to market research and business outreach doubled in three months.
• The video claims AI can complete tasks involving college-level knowledge 12 times faster.
• The presenter says Salesforce reduced its sales workforce from 9,000 to 5,000, attributing the change to automation. This is presented as a claim in the video and is not independently verified here.
• Customer service is described as highly exposed in both Microsoft and Anthropic research. The video also claims 90% of the public may not realize they are interacting with AI in some support situations.

Presenter and sources: Nishant Chahar; Microsoft jobs report and Anthropic jobs report, as described by the presenter.