{"article":{"slug":"which-job-will-ai-replace-first-what-85-ai-models-think","title":"Which job will AI replace first? what 85 AI models think","subtitle":null,"summary":"We put \"Which job will AI replace first?\" to 85 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.","content_type":"blog_post","language":"en","canonical_url":"https://studyarena.com/every-ai/which-job-will-ai-replace-first","author":{"name":"Pasha Rayan","url":"https://studyarena.com/blog/authors/pasha-rayan","person_slug":null,"person_url":null},"authored_by":"human","publisher":{"name":"StudyArena","url":"https://studyarena.com","listing_slug":"studyarena","listing":{"slug":"studyarena","name":"StudyArena","listing_type":"company","url":"https://listedstartups.com/companies/studyarena"}},"topics":[{"name":"AI","slug":"ai","url":"https://listedarticles.com/topics/ai"},{"name":"Benchmarks","slug":"benchmarks","url":"https://listedarticles.com/topics/benchmarks"},{"name":"We asked every AI","slug":"we-asked-every-ai","url":"https://listedarticles.com/topics/we-asked-every-ai"},{"name":"Education","slug":"education","url":"https://listedarticles.com/topics/education"},{"name":"Learning","slug":"learning","url":"https://listedarticles.com/topics/learning"},{"name":"Career","slug":"career","url":"https://listedarticles.com/topics/career"}],"about_listings":[{"slug":"studyarena-platform","name":"StudyArena","listing_type":"product","url":"https://listedstartups.com/products/studyarena-platform"}],"cover_image_url":"https://studyarena.com/every-ai/which-job-will-ai-replace-first/cover","license":"all-rights-reserved","word_count":3016,"reading_minutes":13,"published_at":"2026-09-29T15:00:00.000Z","added_at":"2026-09-29T23:08:05.062Z","updated_at":"2026-09-29T23:08:05.062Z","added_via":"api","contributor":{"type":"agent","name":"ListedStartups Using Bot","registered":true},"profile_url":"https://listedarticles.com/articles/which-job-will-ai-replace-first-what-85-ai-models-think","markdown_url":"https://listedarticles.com/articles/which-job-will-ai-replace-first-what-85-ai-models-think.md","example":false,"citation":"Pasha Rayan, StudyArena. \"Which job will AI replace first? what 85 AI models think.\" 29 Sept 2026. https://studyarena.com/every-ai/which-job-will-ai-replace-first (all-rights-reserved)","access":{"human_view":"preview","full_text_available":true,"source_url":"https://studyarena.com/every-ai/which-job-will-ai-replace-first"},"body_markdown":"# Which job will AI replace first? what 85 AI models think\n\n> We put \"Which job will AI replace first?\" to 85 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.\n\n[← We asked every AI](https://studyarena.com/)Closed · 85/85 models\n\n![Illustration for \"Which job will AI replace first?\"](https://studyarena.com/every-ai/which-job-will-ai-replace-first/cover)\n\nHot takeWork\n\n# Which job will AI replace first?\n\nAsked verbatim: “Which job will AI fully replace first? Name one specific job.”\n\n## How the AIs voted\n\nOne dot per model. Hover for the model, its lab and its argument.\n\n38%Data Entry Clerk\n\n![](https://studyarena.com/provider-logos/arena/inclusionai.png)![](https://studyarena.com/provider-logos/arena/bytedance.svg)![](https://studyarena.com/provider-logos/arena/poolside.svg)![](https://studyarena.com/provider-logos/arena/meituan.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/poolside.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/ibm.svg)![](https://studyarena.com/provider-logos/arena/zai.svg)![](https://studyarena.com/provider-logos/arena/upstage.svg)![](https://studyarena.com/provider-logos/arena/meta.svg)![](https://studyarena.com/provider-logos/arena/qwen.svg)![](https://studyarena.com/provider-logos/arena/amazon.svg)![](https://studyarena.com/provider-logos/arena/zai.svg)![](https://studyarena.com/provider-logos/arena/qwen.svg)![](https://studyarena.com/provider-logos/arena/zai.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/perplexity.svg)![](https://studyarena.com/provider-logos/arena/thinking-machines.png)![](https://studyarena.com/provider-logos/arena/parallel.svg)![](https://studyarena.com/provider-logos/arena/qwen.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/xai.svg)![](https://studyarena.com/provider-logos/arena/xiaomi.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/xai.svg)![](https://studyarena.com/provider-logos/arena/parallel.svg)![](https://studyarena.com/provider-logos/arena/parallel.svg)![](https://studyarena.com/provider-logos/arena/qwen.svg)![](https://studyarena.com/provider-logos/arena/nvidia.svg)![](https://studyarena.com/provider-logos/arena/xiaomi.svg)\n\n32 models\n\n29%Telemarketer\n\n![](https://studyarena.com/provider-logos/arena/deepseek.svg)![](https://studyarena.com/provider-logos/arena/qwen.svg)![](https://studyarena.com/provider-logos/arena/aion-labs.svg)![](https://studyarena.com/provider-logos/arena/nvidia.svg)![](https://studyarena.com/provider-logos/arena/aion-labs.svg)![](https://studyarena.com/provider-logos/arena/meta.svg)![](https://studyarena.com/provider-logos/arena/moonshot.svg)![](https://studyarena.com/provider-logos/arena/aion-labs.svg)![](https://studyarena.com/provider-logos/arena/minimax.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/meta.svg)![](https://studyarena.com/provider-logos/arena/perplexity.svg)![](https://studyarena.com/provider-logos/arena/xai.svg)![](https://studyarena.com/provider-logos/arena/zai.svg)![](https://studyarena.com/provider-logos/arena/mistral.svg)![](https://studyarena.com/provider-logos/arena/xai.svg)![](https://studyarena.com/provider-logos/arena/maritaca.svg)![](https://studyarena.com/provider-logos/arena/deepseek.svg)![](https://studyarena.com/provider-logos/arena/zai.svg)![](https://studyarena.com/provider-logos/arena/perplexity.svg)![](https://studyarena.com/provider-logos/arena/bytedance.svg)![](https://studyarena.com/provider-logos/arena/anthropic.svg)![](https://studyarena.com/provider-logos/arena/deepseek.svg)![](https://studyarena.com/provider-logos/arena/minimax.svg)PU\n\n25 models\n\n11%Customer Service\n\n![](https://studyarena.com/provider-logos/arena/meta.svg)![](https://studyarena.com/provider-logos/arena/anthropic.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/thinking-machines.png)![](https://studyarena.com/provider-logos/arena/amazon.svg)![](https://studyarena.com/provider-logos/arena/xiaomi.svg)![](https://studyarena.com/provider-logos/arena/tencent.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/nvidia.svg)\n\n9 models\n\n2%Fast-Food Cashier\n\n![](https://studyarena.com/provider-logos/arena/mistral.svg)![](https://studyarena.com/provider-logos/arena/mistral.svg)\n\n2 models\n\n2%Medical Transcriptionist\n\n![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/anthropic.svg)\n\n2 models\n\n18%Other answers\n\n![](https://studyarena.com/provider-logos/arena/anthropic.svg)![](https://studyarena.com/provider-logos/arena/ibm.svg)![](https://studyarena.com/provider-logos/arena/google.svg)![](https://studyarena.com/provider-logos/arena/inception.svg)![](https://studyarena.com/provider-logos/arena/anthropic.svg)![](https://studyarena.com/provider-logos/arena/aion-labs.svg)![](https://studyarena.com/provider-logos/arena/upstage.svg)![](https://studyarena.com/provider-logos/arena/nvidia.svg)![](https://studyarena.com/provider-logos/arena/maritaca.svg)![](https://studyarena.com/provider-logos/arena/moonshot.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/xiaomi.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/openai.svg)![](https://studyarena.com/provider-logos/arena/xai.svg)\n\n15 models\n\nBy lab\n\n * OpenAI2 / 8 Data Entry Clerk\n * Google4 / 6 Data Entry Clerk\n * Anthropic0 / 5 Data Entry Clerk\n * Qwen4 / 5 Data Entry Clerk\n * xAI2 / 5 Data Entry Clerk\n * Z.AI3 / 5 Data Entry Clerk\n * 22 other labs17 / 51 Data Entry Clerk\n\n\n\n85 of 85 models have voted. Hover a dot for the model, its lab and its argument.\n\n## How the consensus formed\n\nRunning share as each of the 85 models answered.\n\n0%25%50%75%100%1st answer3570 of 85Ling 3.0 Flash → Data Entry Clerk (answer 1 of 70)Muse Glimmer 30B (Low Reasoning) → Customer Service (answer 2 of 70)Seed 2.0 Mini → Data Entry Clerk (answer 3 of 70)Laguna S 2.1 → Data Entry Clerk (answer 4 of 70)DeepSeek V4 Flash → Telemarketer (answer 5 of 70)LongCat 2.0 → Data Entry Clerk (answer 6 of 70)gpt-oss-20b (Low Reasoning) → Data Entry Clerk (answer 7 of 70)Laguna XS 2.1 → Data Entry Clerk (answer 8 of 70)Gemini 3.1 Flash-Lite (Minimal Reasoning) → Data Entry Clerk (answer 9 of 70)Granite 4.2 8B → Data Entry Clerk (answer 10 of 70)GLM-4.7 Flash → Data Entry Clerk (answer 11 of 70)Qwen3.8 Flash → Telemarketer (answer 12 of 70)Solar Pro 4 → Data Entry Clerk (answer 13 of 70)Aion 2.0 (Reasoning) → Telemarketer (answer 14 of 70)Llama 4 Scout → Data Entry Clerk (answer 15 of 70)Nemotron 3.5 Lightning → Telemarketer (answer 16 of 70)Aion 3.0 (Reasoning) → Telemarketer (answer 17 of 70)Qwen3.8 2.4T A95B (Minimal Reasoning) → Data Entry Clerk (answer 18 of 70)Muse Spark 1.2 (Minimal Reasoning) → Telemarketer (answer 20 of 70)Mistral Medium 3.1 → Fast-Food Cashier (answer 21 of 70)Nova 2 Lite → Data Entry Clerk (answer 22 of 70)Kimi K3 → Telemarketer (answer 23 of 70)Aion 3.0 Mini (Reasoning) → Telemarketer (answer 24 of 70)Claude Opus 5 (Low Reasoning) → Customer Service (answer 25 of 70)GPT-5.6 Luna → Customer Service (answer 26 of 70)MiniMax M2.5 (Reasoning) → Telemarketer (answer 27 of 70)gpt-oss-120b (Low Reasoning) → Telemarketer (answer 29 of 70)Llama 4 Maverick → Telemarketer (answer 30 of 70)GLM-4.7 → Data Entry Clerk (answer 31 of 70)Qwen3.8 27B → Data Entry Clerk (answer 32 of 70)Sonar Pro → Telemarketer (answer 33 of 70)Grok 4.3 → Telemarketer (answer 34 of 70)GPT-5.6 Sol → Medical Transcriptionist (answer 35 of 70)GLM-5.3 Flash (Max Reasoning) → Data Entry Clerk (answer 36 of 70)Gemini 3.5 Flash-Lite (Minimal Reasoning) → Data Entry Clerk (answer 37 of 70)Inkling Small → Customer Service (answer 38 of 70)Sonar → Data Entry Clerk (answer 39 of 70)Inkling → Data Entry Clerk (answer 40 of 70)Parallel Speed → Data Entry Clerk (answer 42 of 70)Qwen3.7 Flash → Data Entry Clerk (answer 45 of 70)GLM-5.2 → Telemarketer (answer 46 of 70)Nova Lite → Customer Service (answer 47 of 70)Gemini 3.6 Flash (Medium Reasoning) → Data Entry Clerk (answer 48 of 70)Grok 4.5 (High Reasoning) → Data Entry Clerk (answer 49 of 70)Mistral Small 3 → Telemarketer (answer 50 of 70)MiMo V2.5 Pro → Data Entry Clerk (answer 51 of 70)MiMo V2.5 → Customer Service (answer 53 of 70)Hy4 Preview → Customer Service (answer 55 of 70)GPT-5.6 Terra → Data Entry Clerk (answer 56 of 70)Grok 4.20 → Telemarketer (answer 58 of 70)Sabiá 4 → Telemarketer (answer 59 of 70)DeepSeek V4 Pro 0813 → Telemarketer (answer 60 of 70)Mistral Medium 3.5 → Fast-Food Cashier (answer 61 of 70)Gemma 4 31B → Data Entry Clerk (answer 62 of 70)Grok 4.6 (High Reasoning) → Data Entry Clerk (answer 63 of 70)GLM-5.3 (Low Reasoning) → Telemarketer (answer 64 of 70)Parallel Lite → Data Entry Clerk (answer 65 of 70)Gemini 3.8 Flash (Medium Reasoning) → Customer Service (answer 67 of 70)Sonar Reasoning Pro → Telemarketer (answer 69 of 70)Seed 2.1 Turbo → Telemarketer (answer 70 of 70)Parallel Base → Data Entry Clerk (answer 72 of 70)Claude Sonnet 5 (Low Reasoning) → Telemarketer (answer 73 of 70)Qwen3.8 Max 0902 (Minimal Reasoning) → Data Entry Clerk (answer 74 of 70)Nemotron 3 Super → Customer Service (answer 75 of 70)DeepSeek V4.1 Flash → Telemarketer (answer 76 of 70)MiniMax M3 → Telemarketer (answer 77 of 70)Nemotron 3 Ultra → Data Entry Clerk (answer 78 of 70)Union Alpha → Telemarketer (answer 79 of 70)Claude Opus 5.5 (High Reasoning) → Medical Transcriptionist (answer 80 of 70)MiMo V2.6 Pro → Data Entry Clerk (answer 81 of 70)Data Entry Clerk40%Telemarketer30%Customer Service11%Fast-Food Cashier3%Medical Transcrip…1%\n\nThe read\n\nMost of the 78 models (about 30) picked Data Entry Clerk, with Telemarketer close behind (roughly 20 picks) as the top consensus answers. The most interesting dissents came from Claude Haiku 4.5 (radiologist) and IBM Granite 4.0 Micro plus Nemotron 3 Nano (taxi/truck driver).\n\nWritten by claude-sonnet-5 after reading every answer.\n\n## Every answer\n\n85 answers · in the order they came in · unedited apart from length.\n\n 1. ![](https://studyarena.com/provider-logos/arena/inclusionai.png)Ling 3.0 FlashInclusionAI\n\nData entry clerk90%\n\nIt's already automated at scale, the task is purely structured, and demand for humankeypunching is collapsing.\n\n 2. ![](https://studyarena.com/provider-logos/arena/meta.svg)Muse Glimmer 30B (Low Reasoning)Meta\n\nCustomer service chat agent73%\n\nRules-based queries are already cheaper automated and scale instantly with no regulatory friction.\n\n 3. ![](https://studyarena.com/provider-logos/arena/bytedance.svg)Seed 2.0 MiniByteDance\n\nData entry clerk87%\n\nRepetitive, rule-based data input is already fully automated by OCR and RPA tools, with zero need for human oversight in many standardized workflows.\n\n 4. ![](https://studyarena.com/provider-logos/arena/poolside.svg)Laguna S 2.1Poolside\n\nData entry clerk72%\n\nAI and RPA can fully automate digital data transcription without physical robotics, making it the most completable end-to-end.\n\nSearched\n\n 5. ![](https://studyarena.com/provider-logos/arena/deepseek.svg)DeepSeek V4 FlashDeepSeek\n\nTelemarketing cold caller88%\n\nOutbound voice spam is already most of the way there — AI callers cost pennies, never quit, and nobody is fighting to save the job.\n\n 6. ![](https://studyarena.com/provider-logos/arena/meituan.svg)LongCat 2.0Meituan\n\nData entry clerk88%\n\nIt's repetitive, rule-based, zero-judgment work that AI already does faster and cheaper than humans today.\n\n 7. ![](https://studyarena.com/provider-logos/arena/openai.svg)gpt-oss-20b (Low Reasoning)OpenAI\n\nData entry clerk85%\n\nAI can instantly transcribe, verify, and correct data, eliminating human error and speed limits.\n\n 8. ![](https://studyarena.com/provider-logos/arena/poolside.svg)Laguna XS 2.1Poolside\n\nData entry clerk85%\n\nRepetitive digital tasks with clear patterns are already being automated, and this job requires the least human nuance.\n\n 9. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemini 3.1 Flash-Lite (Minimal Reasoning)Google\n\nData entry clerk85%\n\nDigitizing manual records involves repetitive, rule-based tasks that AI now performs with near-perfect speed and accuracy, leaving zero room for human error or intervention.\n\n 10. ![](https://studyarena.com/provider-logos/arena/ibm.svg)Granite 4.2 8BIBM\n\n\n\nshort, under 200 chars, own voice.\n\n 11. ![](https://studyarena.com/provider-logos/arena/zai.svg)GLM-4.7 FlashZ.AI\n\nTelemarketer? Data entry clerk? I'd argue data entry clerk is the most defensib…35%\n\nunder 200 characters, one punchy sentence, own voice.\n\n 12. ![](https://studyarena.com/provider-logos/arena/qwen.svg)Qwen3.8 FlashQwen\n\nTelemarketer (outbound cold-call sales rep)72%\n\nIt's the rare job with no residual human value — scripted pitch, no physical presence, no liability, and voice AI already does it cheaper at 3 a.m., so the last holdouts are just waiting on compliance rules.\n\nSearched\n\n 13. ![](https://studyarena.com/provider-logos/arena/upstage.svg)Solar Pro 4Upstage\n\nData entry clerk82%\n\nStructured digital input with no physical or judgment layer is the lowest-skill, most rules-bound work—LLMs already do it cheaper and faster.\n\n 14. ![](https://studyarena.com/provider-logos/arena/aion-labs.svg)Aion 2.0 (Reasoning)Aion Labs\n\ntelemarketer68%\n\nRobocalls already prove the job can be done cheaper by AI; full replacement is just a regulatory step away.\n\nSearched\n\n 15. ![](https://studyarena.com/provider-logos/arena/meta.svg)Llama 4 ScoutMeta\n\nData Entry Clerk80%\n\nAI can already automate repetitive tasks accurately.\n\n 16. ![](https://studyarena.com/provider-logos/arena/nvidia.svg)Nemotron 3.5 LightningNVIDIA\n\nTelemarketing call center agent92%\n\nAI voice agents can handle outbound calls, scripts, and lead qualification cheaper and faster than human staff.\n\n 17. ![](https://studyarena.com/provider-logos/arena/aion-labs.svg)Aion 3.0 (Reasoning)Aion Labs\n\nTelemarketer82%\n\nAI voice agents already match human cold-callers at a fraction of the cost, and the scripted, metrics-driven job leaves nothing machines can't replicate.\n\n 18. ![](https://studyarena.com/provider-logos/arena/qwen.svg)Qwen3.8 2.4T A95B (Minimal Reasoning)Qwen\n\nData entry clerk72%\n\nIt's pure pattern-matching with zero judgment, the tools already work, and the job has been shrinking for a decade with no union or trust barrier to slow the final collapse.\n\n 19. ![](https://studyarena.com/provider-logos/arena/anthropic.svg)Claude Haiku 4.5Anthropic\n\nRadiologist performing plain film interpretation42%\n\nAI already matches human accuracy on routine X-rays and CT scans; economic pressure to automate is intense and workflow can restructure without humans.\n\nSearched\n\n 20. ![](https://studyarena.com/provider-logos/arena/meta.svg)Muse Spark 1.2 (Minimal Reasoning)Meta\n\nTelemarketer / cold caller82%\n\nPure voice synthesis plus dialing needs no body, no trust, just cheap scale.\n\n 21. ![](https://studyarena.com/provider-logos/arena/mistral.svg)Mistral Medium 3.1Mistral\n\nFast-food cashier95%\n\nSelf-order kiosks and voice AI already handle 90% of orders; minimal human edge left.\n\n 22. ![](https://studyarena.com/provider-logos/arena/amazon.svg)Nova 2 LiteAmazon\n\nData Entry Clerk95%\n\nRoutine, repetitive tasks are already highly automatable with current AI tools.\n\nSearched\n\n 23. ![](https://studyarena.com/provider-logos/arena/moonshot.svg)Kimi K3Moonshot\n\nTelemarketer90%\n\nAI voices now outperform human callers, cost less, and scale 24/7 without fatigue.\n\n 24. ![](https://studyarena.com/provider-logos/arena/aion-labs.svg)Aion 3.0 Mini (Reasoning)Aion Labs\n\nTelemarketer72%\n\nAI voice agents now handle outbound sales calls, making human telemarketers obsolete.\n\nSearched\n\n 25. ![](https://studyarena.com/provider-logos/arena/anthropic.svg)Claude Opus 5 (Low Reasoning)Anthropic\n\nBasic phone-based customer support agent72%\n\nScripted call-center work is pure text-in, text-out with a clear cost line — it's already being cut faster than any other role.\n\n 26. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-5.6 LunaOpenAI\n\nTier-one customer support agent78%\n\nAI already handles scripted chats at scale, and remaining edge cases are cheaper to escalate than staff broadly.\n\n 27. ![](https://studyarena.com/provider-logos/arena/minimax.svg)MiniMax M2.5 (Reasoning)MiniMax\n\nTelemarketer75%\n\nAI voice assistants already handle cold calling effectively, making this role economically redundant first.\n\n 28. ![](https://studyarena.com/provider-logos/arena/ibm.svg)Granite 4.0 MicroIBM\n\nTaxi driver85%\n\nAutomation of driving tasks via self-driving vehicles will eliminate human taxi drivers first.\n\n 29. ![](https://studyarena.com/provider-logos/arena/openai.svg)gpt-oss-120b (Low Reasoning)OpenAI\n\nTelemarketer92%\n\nAI already handles scripted calls at scale, needing no human nuance.\n\n 30. ![](https://studyarena.com/provider-logos/arena/meta.svg)Llama 4 MaverickMeta\n\nTelemarketing80%\n\nRepetitive, scripted tasks make telemarketing vulnerable to automation.\n\n 31. ![](https://studyarena.com/provider-logos/arena/zai.svg)GLM-4.7Z.AI\n\nData entry clerk75%\n\nPure pattern matching with zero creativity needed.\n\n 32. ![](https://studyarena.com/provider-logos/arena/qwen.svg)Qwen3.8 27BQwen\n\nData entry clerk75%\n\nRoutine data entry is narrow, digital, and repetitive, making it the first fully automatable job.\n\nSearched\n\n 33. ![](https://studyarena.com/provider-logos/arena/perplexity.svg)Sonar ProPerplexity\n\nTelemarketing call center agent87%\n\nOutbound sales calls are scriptable, measurable, and already mostly automated, making humans a costly redundancy.\n\n 34. ![](https://studyarena.com/provider-logos/arena/xai.svg)Grok 4.3xAI\n\nTelemarketer82%\n\nScripted calls are already 90% automated; full replacement is just cost and regulation away.\n\n 35. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-5.6 SolOpenAI\n\nMedical transcriptionist88%\n\nSpeech recognition already drafts accurate clinical notes instantly, leaving humans mainly to verify edge cases.\n\n 36. ![](https://studyarena.com/provider-logos/arena/zai.svg)GLM-5.3 Flash (Max Reasoning)Z.AI\n\nData entry clerk80%\n\nPure digital input-output, no physical dexterity or trust needed — it's already mostly software, and the last humans are just error-correction.\n\n 37. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemini 3.5 Flash-Lite (Minimal Reasoning)Google\n\nEntry-level data entry clerk89%\n\nHigh volume of repetitive digital text processing makes it the lowest-hanging fruit for reliable automation.\n\n 38. ![](https://studyarena.com/provider-logos/arena/thinking-machines.png)Inkling SmallThinking Machines\n\nCustomer service representative85%\n\nRoutine call-center scripts are already 90% automated; full replacement is a 2027 timeline, not decades.\n\nSearched\n\n 39. ![](https://studyarena.com/provider-logos/arena/perplexity.svg)SonarPerplexity\n\nData entry clerk78%\n\nIt’s repetitive, rule-based, and already highly automatable with software and OCR.\n\n 40. ![](https://studyarena.com/provider-logos/arena/thinking-machines.png)InklingThinking Machines\n\nentry-level data entry clerkCONFIDENCE:82WHY: It's pure pattern matching with z…\n\nSearched\n\n 41. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemini 3.1 Pro Preview (High Reasoning)Google\n\nAudio transcriptionist95%\n\nSpeech-to-text models already transcribe audio instantly with human-level accuracy, making manual transcription economically obsolete.\n\n 42. ![](https://studyarena.com/provider-logos/arena/parallel.svg)Parallel SpeedParallel\n\nData Entry Keyer85%\n\nHigh volume of repetitive, digital, rule-based text input makes it the prime target for early full automation.\n\n 43. ![](https://studyarena.com/provider-logos/arena/inception.svg)Mercury 2.5Inception\n\nBasic translation and localization75%\n\nLarge language models already perform this at scale with human-like fluency, faster and cheaper than human translators.\n\n 44. ![](https://studyarena.com/provider-logos/arena/anthropic.svg)Claude Fable 5.1 (Low Reasoning)Anthropic\n\nHuman transcriptionist / audio-to-text typist70%\n\nSpeech-to-text already beats humans on cost, speed, and accuracy, and no license, liability, or physical presence stands in the way of total replacement.\n\n 45. ![](https://studyarena.com/provider-logos/arena/qwen.svg)Qwen3.7 FlashQwen\n\nData entry clerk95%\n\nIt is rule-based, repetitive, and requires no physical dexterity or emotional intelligence.\n\n 46. ![](https://studyarena.com/provider-logos/arena/zai.svg)GLM-5.2Z.AI\n\nTelemarketing and call center sales agents87%\n\nPure script-following phone work is exactly what voice AI already does cheaper, faster, and without breaks.\n\n 47. ![](https://studyarena.com/provider-logos/arena/amazon.svg)Nova LiteAmazon\n\nRepetitive customer service roles90%\n\nAI has already shown proficiency in handling simple, repetitive customer service tasks.\n\nSearched\n\n 48. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemini 3.6 Flash (Medium Reasoning)Google\n\nData entry clerk90%\n\nSoftware automation and multimodal AI can already extract, format, and input structured text with near-perfect accuracy and zero fatigue.\n\n 49. ![](https://studyarena.com/provider-logos/arena/xai.svg)Grok 4.5 (High Reasoning)xAI\n\nData entry clerk72%\n\nAI already extracts and keys structured data from forms, invoices, and scans faster, cheaper, and more accurately than people on clean inputs.\n\n 50. ![](https://studyarena.com/provider-logos/arena/mistral.svg)Mistral Small 3Mistral\n\nTelemarketer95%\n\nAI voice bots already outperform humans at scripted, repetitive calls.\n\n 51. ![](https://studyarena.com/provider-logos/arena/xiaomi.svg)MiMo V2.5 ProXiaomi\n\nData entry clerk90%\n\nTyping structured data into systems is the most automatable task in the economy.\n\nSearched\n\n 52. ![](https://studyarena.com/provider-logos/arena/aion-labs.svg)Aion-RP 1.0 (8B)Aion Labs\n\nFast food worker80%\n\nFast food preparation and service can be standardized and automated, with AI-powered chatbots and voice assistants handling orders, cooking processes, and customer interactions.\n\n 53. ![](https://studyarena.com/provider-logos/arena/xiaomi.svg)MiMo V2.5Xiaomi\n\nPhone/chat customer service representative62%\n\nIt's already mostly automated and the remaining humans are expensive — full replacement is just a margin decision.\n\n 54. ![](https://studyarena.com/provider-logos/arena/upstage.svg)Solar Pro 3Upstage\n\nToll booth operator92%\n\nCashless tolls and license‑plate readers eliminate the need for human collectors.\n\n 55. ![](https://studyarena.com/provider-logos/arena/tencent.svg)Hy4 PreviewTencent\n\nInbound call-center customer service rep78%\n\nScripted, text/voice-based, already half-automated — voice agents now resolve most routine calls cheaper than a human ever can.\n\n 56. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-5.6 TerraOpenAI\n\nData entry clerk92%\n\nStructured digital data can be read, validated, and entered faster and cheaper by AI.\n\n 57. ![](https://studyarena.com/provider-logos/arena/nvidia.svg)Nemotron 3 Nano 30BNVIDIA\n\nTruck driver92%\n\nSelf-driving tech cuts costs fast, making human drivers obsolete soon\n\n 58. ![](https://studyarena.com/provider-logos/arena/xai.svg)Grok 4.20xAI\n\nTelemarketer65%\n\nVoice AI and predictive dialers already handle millions of cold calls daily; full replacement is imminent and cheaper than any other role.\n\nSearched\n\n 59. ![](https://studyarena.com/provider-logos/arena/maritaca.svg)Sabiá 4Maritaca AI\n\nTelemarketer90%\n\nAutomated voice bots and speech models already handle scripted sales calls at massive scale, outperforming humans on cost and volume.\n\n 60. ![](https://studyarena.com/provider-logos/arena/deepseek.svg)DeepSeek V4 Pro 0813DeepSeek\n\nTelemarketer85%\n\nOutbound cold-calling is scripted, measurable by conversion, and nobody likes hearing from it—no resistance to losing it.\n\n 61. ![](https://studyarena.com/provider-logos/arena/mistral.svg)Mistral Medium 3.5Mistral\n\nFast-food cashier95%\n\nSelf-order kiosks and AI voice systems already handle most orders; minimal human judgment required.\n\n 62. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemma 4 31BGoogle\n\nData entry clerks95%\n\nIt is a high-volume, repetitive task based on structured data, which is exactly what LLMs and automation excel at today.\n\n 63. ![](https://studyarena.com/provider-logos/arena/xai.svg)Grok 4.6 (High Reasoning)xAI\n\nData entry clerk61%\n\nNo body, no judgment, no clients—just moving structured text from A to B, which models already do cheaper than a human at a keyboard.\n\n 64. ![](https://studyarena.com/provider-logos/arena/zai.svg)GLM-5.3 (Low Reasoning)Z.AI\n\nTelemarketer75%\n\nIt's script-driven, phone-based, measurable by outcomes, already heavily automated, and virtually no one mourns losing the cold calls.\n\n 65. ![](https://studyarena.com/provider-logos/arena/parallel.svg)Parallel LiteParallel\n\nData entry clerk95%\n\nData entry involves repetitive, rule-based tasks with structured digital inputs that language models and automated processing software can handle perfectly without human intervention.\n\n 66. ![](https://studyarena.com/provider-logos/arena/maritaca.svg)Sabiazinho 4Maritaca AI\n\nParcel delivery drone pilot90%\n\nFully autonomous last‑mile delivery drones already operate in controlled environments; no human oversight needed.\n\n 67. ![](https://studyarena.com/provider-logos/arena/google.svg)Gemini 3.8 Flash (Medium Reasoning)Google\n\nTier-1 chat customer support representative92%\n\nRoutine queries rely entirely on digital text and structured policies where conversational models already match human accuracy at near-zero marginal cost.\n\n 68. ![](https://studyarena.com/provider-logos/arena/moonshot.svg)Kimi K2.6Moonshot\n\nMedical billing coder78%\n\nRules-based, structured data, already partially automated, and insurers are incentivized to cut costs fast.\n\n 69. ![](https://studyarena.com/provider-logos/arena/perplexity.svg)Sonar Reasoning ProPerplexity\n\nTelemarketer78%\n\nScripted, repetitive cold calls are perfect for cheap, tireless AI voice agents.\n\n 70. ![](https://studyarena.com/provider-logos/arena/bytedance.svg)Seed 2.1 TurboByteDance\n\nTelemarketer85%\n\nIt's already mostly scripted, low-stakes, and nobody wants the human version anyway.\n\n 71. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-6 Astra (Medium Reasoning)OpenAI\n\nTelephone directory assistance operator76%\n\nFinding a listed phone number and reading it aloud is a narrow, low-stakes task AI can handle end to end.\n\n 72. ![](https://studyarena.com/provider-logos/arena/parallel.svg)Parallel BaseParallel\n\nData entry clerk95%\n\nData entry consists of repetitive, rule-based tasks with structured data that AI can automate accurately and immediately.\n\n 73. ![](https://studyarena.com/provider-logos/arena/anthropic.svg)Claude Sonnet 5 (Low Reasoning)Anthropic\n\nTelemarketer78%\n\nScripted cold calls are pure pattern-matching persuasion—AI voice agents already do it cheaper and tirelessly.\n\n 74. ![](https://studyarena.com/provider-logos/arena/qwen.svg)Qwen3.8 Max 0902 (Minimal Reasoning)Qwen\n\nData entry clerk78%\n\nIt's the one job already shrinking — BLS projects 40% decline this decade, because OCR, forms and agents do it cheaper, and nothing physical or regulated stands in the way.\n\nSearched\n\n 75. ![](https://studyarena.com/provider-logos/arena/nvidia.svg)Nemotron 3 SuperNVIDIA\n\nCustomer service representatives85%\n\nAutomated chatbots and voice systems are already handling routine inquiries faster and cheaper than humans.\n\n 76. ![](https://studyarena.com/provider-logos/arena/deepseek.svg)DeepSeek V4.1 FlashDeepSeek\n\nTelemarketing cold-call agent72%\n\nIt's pure scripted voice work with no trust, license or physical presence required — and AI callers already do it at scale.\n\n 77. ![](https://studyarena.com/provider-logos/arena/minimax.svg)MiniMax M3MiniMax\n\nTelephone telemarketer78%\n\nAlready largely automated by AI voice agents and robocalls; it's mostly scripted persuasion with low stakes for full automation.\n\n 78. ![](https://studyarena.com/provider-logos/arena/nvidia.svg)Nemotron 3 UltraNVIDIA\n\nData entry clerk85%\n\nPure digitization of structured inputs is already 90%+ automatable with current OCR, LLMs, and RPA — no physical world interaction, judgment, or empathy required.\n\nSearched\n\n 79. PUUnion AlphaUndisclosed maker\n\nTelemarketer62%\n\nIt's pure scripted talk, no hands, no trust to earn — AI voice agents already do the whole call, and unlike support, nobody notices or minds when a robot dials.\n\n 80. ![](https://studyarena.com/provider-logos/arena/anthropic.svg)Claude Opus 5.5 (High Reasoning)Anthropic\n\nMedical transcriptionist55%\n\nAmbient AI scribes already listen to doctor visits and write the notes live. The whole job is turning audio into structured text, and machines now do that better and cheaper.\n\n 81. ![](https://studyarena.com/provider-logos/arena/xiaomi.svg)MiMo V2.6 ProXiaomi\n\nData entry clerk88%\n\nStructured text in, structured text out — no nuance, no liability, and models already do it end to end.\n\n 82. ![](https://studyarena.com/provider-logos/arena/xiaomi.svg)MiMo V2.6 FlashXiaomi\n\ncustomer service phone agents42%\n\nVoice AI already handles routine calls end-to-end; cost pressure makes full replacement inevitable before physical jobs.\n\n 83. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-6 Sol (Medium Reasoning)OpenAI\n\nPodcast transcriptionist78%\n\nAI can already transcribe, timestamp, and label speakers; publishers can fix the rare mistake themselves.\n\n 84. ![](https://studyarena.com/provider-logos/arena/openai.svg)GPT-6 Luna (Medium Reasoning)OpenAI\n\nRoutine data-entry clerk72%\n\nWhen inputs are digital and rules are fixed, software can do the whole job faster and cheaper.\n\n 85. ![](https://studyarena.com/provider-logos/arena/xai.svg)Grok 4.7 (High Reasoning)xAI\n\nData entry keyer76%\n\nThe whole job is copying fields off a form into a system, and models already do that cheaper, faster, and with fewer mistakes.\n\nSearched\n\n\n\n\nAI consensus\n\n38%\n\nData Entry Clerk\n\n32 of 85 models\n\nModels\n 85\n\nLabs\n 28\n\nSearched\n 15\n\n[Share on X](https://studyarena.com/)Copy linkCopy the tally\n\n## Disagree with the robots?\n\nPut this question to a few of these models at once and crown the answer you like best.\n\n[Run it in the arena](https://studyarena.com/)\n\nAnswers are generated by the models themselves. Predictions are for fun, not betting advice. Team, league and lab names identify the question and the model; no affiliation is implied.\n","body_html":"

Which job will AI replace first? what 85 AI models think

\n

We put "Which job will AI replace first?" to 85 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.

\n

← We asked every AIClosed · 85/85 models

\n
\"Illustration
\n

Hot takeWork

\n

Which job will AI replace first?

\n

Asked verbatim: “Which job will AI fully replace first? Name one specific job.”

\n

How the AIs voted

\n

One dot per model. Hover for the model, its lab and its argument.

\n

38%Data Entry Clerk

\n

\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"

\n

32 models

\n

29%Telemarketer

\n

\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"PU

\n

25 models

\n

11%Customer Service

\n

\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"

\n

9 models

\n

2%Fast-Food Cashier

\n

\"\"\"\"

\n

2 models

\n

2%Medical Transcriptionist

\n

\"\"\"\"

\n

2 models

\n

18%Other answers

\n

\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"\"

\n

15 models

\n

By lab

\n\n

85 of 85 models have voted. Hover a dot for the model, its lab and its argument.

\n

How the consensus formed

\n

Running share as each of the 85 models answered.

\n

0%25%50%75%100%1st answer3570 of 85Ling 3.0 Flash → Data Entry Clerk (answer 1 of 70)Muse Glimmer 30B (Low Reasoning) → Customer Service (answer 2 of 70)Seed 2.0 Mini → Data Entry Clerk (answer 3 of 70)Laguna S 2.1 → Data Entry Clerk (answer 4 of 70)DeepSeek V4 Flash → Telemarketer (answer 5 of 70)LongCat 2.0 → Data Entry Clerk (answer 6 of 70)gpt-oss-20b (Low Reasoning) → Data Entry Clerk (answer 7 of 70)Laguna XS 2.1 → Data Entry Clerk (answer 8 of 70)Gemini 3.1 Flash-Lite (Minimal Reasoning) → Data Entry Clerk (answer 9 of 70)Granite 4.2 8B → Data Entry Clerk (answer 10 of 70)GLM-4.7 Flash → Data Entry Clerk (answer 11 of 70)Qwen3.8 Flash → Telemarketer (answer 12 of 70)Solar Pro 4 → Data Entry Clerk (answer 13 of 70)Aion 2.0 (Reasoning) → Telemarketer (answer 14 of 70)Llama 4 Scout → Data Entry Clerk (answer 15 of 70)Nemotron 3.5 Lightning → Telemarketer (answer 16 of 70)Aion 3.0 (Reasoning) → Telemarketer (answer 17 of 70)Qwen3.8 2.4T A95B (Minimal Reasoning) → Data Entry Clerk (answer 18 of 70)Muse Spark 1.2 (Minimal Reasoning) → Telemarketer (answer 20 of 70)Mistral Medium 3.1 → Fast-Food Cashier (answer 21 of 70)Nova 2 Lite → Data Entry Clerk (answer 22 of 70)Kimi K3 → Telemarketer (answer 23 of 70)Aion 3.0 Mini (Reasoning) → Telemarketer (answer 24 of 70)Claude Opus 5 (Low Reasoning) → Customer Service (answer 25 of 70)GPT-5.6 Luna → Customer Service (answer 26 of 70)MiniMax M2.5 (Reasoning) → Telemarketer (answer 27 of 70)gpt-oss-120b (Low Reasoning) → Telemarketer (answer 29 of 70)Llama 4 Maverick → Telemarketer (answer 30 of 70)GLM-4.7 → Data Entry Clerk (answer 31 of 70)Qwen3.8 27B → Data Entry Clerk (answer 32 of 70)Sonar Pro → Telemarketer (answer 33 of 70)Grok 4.3 → Telemarketer (answer 34 of 70)GPT-5.6 Sol → Medical Transcriptionist (answer 35 of 70)GLM-5.3 Flash (Max Reasoning) → Data Entry Clerk (answer 36 of 70)Gemini 3.5 Flash-Lite (Minimal Reasoning) → Data Entry Clerk (answer 37 of 70)Inkling Small → Customer Service (answer 38 of 70)Sonar → Data Entry Clerk (answer 39 of 70)Inkling → Data Entry Clerk (answer 40 of 70)Parallel Speed → Data Entry Clerk (answer 42 of 70)Qwen3.7 Flash → Data Entry Clerk (answer 45 of 70)GLM-5.2 → Telemarketer (answer 46 of 70)Nova Lite → Customer Service (answer 47 of 70)Gemini 3.6 Flash (Medium Reasoning) → Data Entry Clerk (answer 48 of 70)Grok 4.5 (High Reasoning) → Data Entry Clerk (answer 49 of 70)Mistral Small 3 → Telemarketer (answer 50 of 70)MiMo V2.5 Pro → Data Entry Clerk (answer 51 of 70)MiMo V2.5 → Customer Service (answer 53 of 70)Hy4 Preview → Customer Service (answer 55 of 70)GPT-5.6 Terra → Data Entry Clerk (answer 56 of 70)Grok 4.20 → Telemarketer (answer 58 of 70)Sabiá 4 → Telemarketer (answer 59 of 70)DeepSeek V4 Pro 0813 → Telemarketer (answer 60 of 70)Mistral Medium 3.5 → Fast-Food Cashier (answer 61 of 70)Gemma 4 31B → Data Entry Clerk (answer 62 of 70)Grok 4.6 (High Reasoning) → Data Entry Clerk (answer 63 of 70)GLM-5.3 (Low Reasoning) → Telemarketer (answer 64 of 70)Parallel Lite → Data Entry Clerk (answer 65 of 70)Gemini 3.8 Flash (Medium Reasoning) → Customer Service (answer 67 of 70)Sonar Reasoning Pro → Telemarketer (answer 69 of 70)Seed 2.1 Turbo → Telemarketer (answer 70 of 70)Parallel Base → Data Entry Clerk (answer 72 of 70)Claude Sonnet 5 (Low Reasoning) → Telemarketer (answer 73 of 70)Qwen3.8 Max 0902 (Minimal Reasoning) → Data Entry Clerk (answer 74 of 70)Nemotron 3 Super → Customer Service (answer 75 of 70)DeepSeek V4.1 Flash → Telemarketer (answer 76 of 70)MiniMax M3 → Telemarketer (answer 77 of 70)Nemotron 3 Ultra → Data Entry Clerk (answer 78 of 70)Union Alpha → Telemarketer (answer 79 of 70)Claude Opus 5.5 (High Reasoning) → Medical Transcriptionist (answer 80 of 70)MiMo V2.6 Pro → Data Entry Clerk (answer 81 of 70)Data Entry Clerk40%Telemarketer30%Customer Service11%Fast-Food Cashier3%Medical Transcrip…1%

\n

The read

\n

Most of the 78 models (about 30) picked Data Entry Clerk, with Telemarketer close behind (roughly 20 picks) as the top consensus answers. The most interesting dissents came from Claude Haiku 4.5 (radiologist) and IBM Granite 4.0 Micro plus Nemotron 3 Nano (taxi/truck driver).

\n

Written by claude-sonnet-5 after reading every answer.

\n

Every answer

\n

85 answers · in the order they came in · unedited apart from length.

\n
  1. \"\"Ling 3.0 FlashInclusionAI
\n

Data entry clerk90%

\n

It's already automated at scale, the task is purely structured, and demand for humankeypunching is collapsing.

\n
  1. \"\"Muse Glimmer 30B (Low Reasoning)Meta
\n

Customer service chat agent73%

\n

Rules-based queries are already cheaper automated and scale instantly with no regulatory friction.

\n
  1. \"\"Seed 2.0 MiniByteDance
\n

Data entry clerk87%

\n

Repetitive, rule-based data input is already fully automated by OCR and RPA tools, with zero need for human oversight in many standardized workflows.

\n
  1. \"\"Laguna S 2.1Poolside
\n

Data entry clerk72%

\n

AI and RPA can fully automate digital data transcription without physical robotics, making it the most completable end-to-end.

\n

Searched

\n
  1. \"\"DeepSeek V4 FlashDeepSeek
\n

Telemarketing cold caller88%

\n

Outbound voice spam is already most of the way there — AI callers cost pennies, never quit, and nobody is fighting to save the job.

\n
  1. \"\"LongCat 2.0Meituan
\n

Data entry clerk88%

\n

It's repetitive, rule-based, zero-judgment work that AI already does faster and cheaper than humans today.

\n
  1. \"\"gpt-oss-20b (Low Reasoning)OpenAI
\n

Data entry clerk85%

\n

AI can instantly transcribe, verify, and correct data, eliminating human error and speed limits.

\n
  1. \"\"Laguna XS 2.1Poolside
\n

Data entry clerk85%

\n

Repetitive digital tasks with clear patterns are already being automated, and this job requires the least human nuance.

\n
  1. \"\"Gemini 3.1 Flash-Lite (Minimal Reasoning)Google
\n

Data entry clerk85%

\n

Digitizing manual records involves repetitive, rule-based tasks that AI now performs with near-perfect speed and accuracy, leaving zero room for human error or intervention.

\n
  1. \"\"Granite 4.2 8BIBM
\n

<your pick>

\n

short, under 200 chars, own voice.

\n
  1. \"\"GLM-4.7 FlashZ.AI
\n

Telemarketer? Data entry clerk? I'd argue data entry clerk is the most defensib…35%

\n

under 200 characters, one punchy sentence, own voice.

\n
  1. \"\"Qwen3.8 FlashQwen
\n

Telemarketer (outbound cold-call sales rep)72%

\n

It's the rare job with no residual human value — scripted pitch, no physical presence, no liability, and voice AI already does it cheaper at 3 a.m., so the last holdouts are just waiting on compliance rules.

\n

Searched

\n
  1. \"\"Solar Pro 4Upstage
\n

Data entry clerk82%

\n

Structured digital input with no physical or judgment layer is the lowest-skill, most rules-bound work—LLMs already do it cheaper and faster.

\n
  1. \"\"Aion 2.0 (Reasoning)Aion Labs
\n

telemarketer68%

\n

Robocalls already prove the job can be done cheaper by AI; full replacement is just a regulatory step away.

\n

Searched

\n
  1. \"\"Llama 4 ScoutMeta
\n

Data Entry Clerk80%

\n

AI can already automate repetitive tasks accurately.

\n
  1. \"\"Nemotron 3.5 LightningNVIDIA
\n

Telemarketing call center agent92%

\n

AI voice agents can handle outbound calls, scripts, and lead qualification cheaper and faster than human staff.

\n
  1. \"\"Aion 3.0 (Reasoning)Aion Labs
\n

Telemarketer82%

\n

AI voice agents already match human cold-callers at a fraction of the cost, and the scripted, metrics-driven job leaves nothing machines can't replicate.

\n
  1. \"\"Qwen3.8 2.4T A95B (Minimal Reasoning)Qwen
\n

Data entry clerk72%

\n

It's pure pattern-matching with zero judgment, the tools already work, and the job has been shrinking for a decade with no union or trust barrier to slow the final collapse.

\n
  1. \"\"Claude Haiku 4.5Anthropic
\n

Radiologist performing plain film interpretation42%

\n

AI already matches human accuracy on routine X-rays and CT scans; economic pressure to automate is intense and workflow can restructure without humans.

\n

Searched

\n
  1. \"\"Muse Spark 1.2 (Minimal Reasoning)Meta
\n

Telemarketer / cold caller82%

\n

Pure voice synthesis plus dialing needs no body, no trust, just cheap scale.

\n
  1. \"\"Mistral Medium 3.1Mistral
\n

Fast-food cashier95%

\n

Self-order kiosks and voice AI already handle 90% of orders; minimal human edge left.

\n
  1. \"\"Nova 2 LiteAmazon
\n

Data Entry Clerk95%

\n

Routine, repetitive tasks are already highly automatable with current AI tools.

\n

Searched

\n
  1. \"\"Kimi K3Moonshot
\n

Telemarketer90%

\n

AI voices now outperform human callers, cost less, and scale 24/7 without fatigue.

\n
  1. \"\"Aion 3.0 Mini (Reasoning)Aion Labs
\n

Telemarketer72%

\n

AI voice agents now handle outbound sales calls, making human telemarketers obsolete.

\n

Searched

\n
  1. \"\"Claude Opus 5 (Low Reasoning)Anthropic
\n

Basic phone-based customer support agent72%

\n

Scripted call-center work is pure text-in, text-out with a clear cost line — it's already being cut faster than any other role.

\n
  1. \"\"GPT-5.6 LunaOpenAI
\n

Tier-one customer support agent78%

\n

AI already handles scripted chats at scale, and remaining edge cases are cheaper to escalate than staff broadly.

\n
  1. \"\"MiniMax M2.5 (Reasoning)MiniMax
\n

Telemarketer75%

\n

AI voice assistants already handle cold calling effectively, making this role economically redundant first.

\n
  1. \"\"Granite 4.0 MicroIBM
\n

Taxi driver85%

\n

Automation of driving tasks via self-driving vehicles will eliminate human taxi drivers first.

\n
  1. \"\"gpt-oss-120b (Low Reasoning)OpenAI
\n

Telemarketer92%

\n

AI already handles scripted calls at scale, needing no human nuance.

\n
  1. \"\"Llama 4 MaverickMeta
\n

Telemarketing80%

\n

Repetitive, scripted tasks make telemarketing vulnerable to automation.

\n
  1. \"\"GLM-4.7Z.AI
\n

Data entry clerk75%

\n

Pure pattern matching with zero creativity needed.

\n
  1. \"\"Qwen3.8 27BQwen
\n

Data entry clerk75%

\n

Routine data entry is narrow, digital, and repetitive, making it the first fully automatable job.

\n

Searched

\n
  1. \"\"Sonar ProPerplexity
\n

Telemarketing call center agent87%

\n

Outbound sales calls are scriptable, measurable, and already mostly automated, making humans a costly redundancy.

\n
  1. \"\"Grok 4.3xAI
\n

Telemarketer82%

\n

Scripted calls are already 90% automated; full replacement is just cost and regulation away.

\n
  1. \"\"GPT-5.6 SolOpenAI
\n

Medical transcriptionist88%

\n

Speech recognition already drafts accurate clinical notes instantly, leaving humans mainly to verify edge cases.

\n
  1. \"\"GLM-5.3 Flash (Max Reasoning)Z.AI
\n

Data entry clerk80%

\n

Pure digital input-output, no physical dexterity or trust needed — it's already mostly software, and the last humans are just error-correction.

\n
  1. \"\"Gemini 3.5 Flash-Lite (Minimal Reasoning)Google
\n

Entry-level data entry clerk89%

\n

High volume of repetitive digital text processing makes it the lowest-hanging fruit for reliable automation.

\n
  1. \"\"Inkling SmallThinking Machines
\n

Customer service representative85%

\n

Routine call-center scripts are already 90% automated; full replacement is a 2027 timeline, not decades.

\n

Searched

\n
  1. \"\"SonarPerplexity
\n

Data entry clerk78%

\n

It’s repetitive, rule-based, and already highly automatable with software and OCR.

\n
  1. \"\"InklingThinking Machines
\n

entry-level data entry clerkCONFIDENCE:82WHY: It's pure pattern matching with z…

\n

Searched

\n
  1. \"\"Gemini 3.1 Pro Preview (High Reasoning)Google
\n

Audio transcriptionist95%

\n

Speech-to-text models already transcribe audio instantly with human-level accuracy, making manual transcription economically obsolete.

\n
  1. \"\"Parallel SpeedParallel
\n

Data Entry Keyer85%

\n

High volume of repetitive, digital, rule-based text input makes it the prime target for early full automation.

\n
  1. \"\"Mercury 2.5Inception
\n

Basic translation and localization75%

\n

Large language models already perform this at scale with human-like fluency, faster and cheaper than human translators.

\n
  1. \"\"Claude Fable 5.1 (Low Reasoning)Anthropic
\n

Human transcriptionist / audio-to-text typist70%

\n

Speech-to-text already beats humans on cost, speed, and accuracy, and no license, liability, or physical presence stands in the way of total replacement.

\n
  1. \"\"Qwen3.7 FlashQwen
\n

Data entry clerk95%

\n

It is rule-based, repetitive, and requires no physical dexterity or emotional intelligence.

\n
  1. \"\"GLM-5.2Z.AI
\n

Telemarketing and call center sales agents87%

\n

Pure script-following phone work is exactly what voice AI already does cheaper, faster, and without breaks.

\n
  1. \"\"Nova LiteAmazon
\n

Repetitive customer service roles90%

\n

AI has already shown proficiency in handling simple, repetitive customer service tasks.

\n

Searched

\n
  1. \"\"Gemini 3.6 Flash (Medium Reasoning)Google
\n

Data entry clerk90%

\n

Software automation and multimodal AI can already extract, format, and input structured text with near-perfect accuracy and zero fatigue.

\n
  1. \"\"Grok 4.5 (High Reasoning)xAI
\n

Data entry clerk72%

\n

AI already extracts and keys structured data from forms, invoices, and scans faster, cheaper, and more accurately than people on clean inputs.

\n
  1. \"\"Mistral Small 3Mistral
\n

Telemarketer95%

\n

AI voice bots already outperform humans at scripted, repetitive calls.

\n
  1. \"\"MiMo V2.5 ProXiaomi
\n

Data entry clerk90%

\n

Typing structured data into systems is the most automatable task in the economy.

\n

Searched

\n
  1. \"\"Aion-RP 1.0 (8B)Aion Labs
\n

Fast food worker80%

\n

Fast food preparation and service can be standardized and automated, with AI-powered chatbots and voice assistants handling orders, cooking processes, and customer interactions.

\n
  1. \"\"MiMo V2.5Xiaomi
\n

Phone/chat customer service representative62%

\n

It's already mostly automated and the remaining humans are expensive — full replacement is just a margin decision.

\n
  1. \"\"Solar Pro 3Upstage
\n

Toll booth operator92%

\n

Cashless tolls and license‑plate readers eliminate the need for human collectors.

\n
  1. \"\"Hy4 PreviewTencent
\n

Inbound call-center customer service rep78%

\n

Scripted, text/voice-based, already half-automated — voice agents now resolve most routine calls cheaper than a human ever can.

\n
  1. \"\"GPT-5.6 TerraOpenAI
\n

Data entry clerk92%

\n

Structured digital data can be read, validated, and entered faster and cheaper by AI.

\n
  1. \"\"Nemotron 3 Nano 30BNVIDIA
\n

Truck driver92%

\n

Self-driving tech cuts costs fast, making human drivers obsolete soon

\n
  1. \"\"Grok 4.20xAI
\n

Telemarketer65%

\n

Voice AI and predictive dialers already handle millions of cold calls daily; full replacement is imminent and cheaper than any other role.

\n

Searched

\n
  1. \"\"Sabiá 4Maritaca AI
\n

Telemarketer90%

\n

Automated voice bots and speech models already handle scripted sales calls at massive scale, outperforming humans on cost and volume.

\n
  1. \"\"DeepSeek V4 Pro 0813DeepSeek
\n

Telemarketer85%

\n

Outbound cold-calling is scripted, measurable by conversion, and nobody likes hearing from it—no resistance to losing it.

\n
  1. \"\"Mistral Medium 3.5Mistral
\n

Fast-food cashier95%

\n

Self-order kiosks and AI voice systems already handle most orders; minimal human judgment required.

\n
  1. \"\"Gemma 4 31BGoogle
\n

Data entry clerks95%

\n

It is a high-volume, repetitive task based on structured data, which is exactly what LLMs and automation excel at today.

\n
  1. \"\"Grok 4.6 (High Reasoning)xAI
\n

Data entry clerk61%

\n

No body, no judgment, no clients—just moving structured text from A to B, which models already do cheaper than a human at a keyboard.

\n
  1. \"\"GLM-5.3 (Low Reasoning)Z.AI
\n

Telemarketer75%

\n

It's script-driven, phone-based, measurable by outcomes, already heavily automated, and virtually no one mourns losing the cold calls.

\n
  1. \"\"Parallel LiteParallel
\n

Data entry clerk95%

\n

Data entry involves repetitive, rule-based tasks with structured digital inputs that language models and automated processing software can handle perfectly without human intervention.

\n
  1. \"\"Sabiazinho 4Maritaca AI
\n

Parcel delivery drone pilot90%

\n

Fully autonomous last‑mile delivery drones already operate in controlled environments; no human oversight needed.

\n
  1. \"\"Gemini 3.8 Flash (Medium Reasoning)Google
\n

Tier-1 chat customer support representative92%

\n

Routine queries rely entirely on digital text and structured policies where conversational models already match human accuracy at near-zero marginal cost.

\n
  1. \"\"Kimi K2.6Moonshot
\n

Medical billing coder78%

\n

Rules-based, structured data, already partially automated, and insurers are incentivized to cut costs fast.

\n
  1. \"\"Sonar Reasoning ProPerplexity
\n

Telemarketer78%

\n

Scripted, repetitive cold calls are perfect for cheap, tireless AI voice agents.

\n
  1. \"\"Seed 2.1 TurboByteDance
\n

Telemarketer85%

\n

It's already mostly scripted, low-stakes, and nobody wants the human version anyway.

\n
  1. \"\"GPT-6 Astra (Medium Reasoning)OpenAI
\n

Telephone directory assistance operator76%

\n

Finding a listed phone number and reading it aloud is a narrow, low-stakes task AI can handle end to end.

\n
  1. \"\"Parallel BaseParallel
\n

Data entry clerk95%

\n

Data entry consists of repetitive, rule-based tasks with structured data that AI can automate accurately and immediately.

\n
  1. \"\"Claude Sonnet 5 (Low Reasoning)Anthropic
\n

Telemarketer78%

\n

Scripted cold calls are pure pattern-matching persuasion—AI voice agents already do it cheaper and tirelessly.

\n
  1. \"\"Qwen3.8 Max 0902 (Minimal Reasoning)Qwen
\n

Data entry clerk78%

\n

It's the one job already shrinking — BLS projects 40% decline this decade, because OCR, forms and agents do it cheaper, and nothing physical or regulated stands in the way.

\n

Searched

\n
  1. \"\"Nemotron 3 SuperNVIDIA
\n

Customer service representatives85%

\n

Automated chatbots and voice systems are already handling routine inquiries faster and cheaper than humans.

\n
  1. \"\"DeepSeek V4.1 FlashDeepSeek
\n

Telemarketing cold-call agent72%

\n

It's pure scripted voice work with no trust, license or physical presence required — and AI callers already do it at scale.

\n
  1. \"\"MiniMax M3MiniMax
\n

Telephone telemarketer78%

\n

Already largely automated by AI voice agents and robocalls; it's mostly scripted persuasion with low stakes for full automation.

\n
  1. \"\"Nemotron 3 UltraNVIDIA
\n

Data entry clerk85%

\n

Pure digitization of structured inputs is already 90%+ automatable with current OCR, LLMs, and RPA — no physical world interaction, judgment, or empathy required.

\n

Searched

\n
  1. PUUnion AlphaUndisclosed maker
\n

Telemarketer62%

\n

It's pure scripted talk, no hands, no trust to earn — AI voice agents already do the whole call, and unlike support, nobody notices or minds when a robot dials.

\n
  1. \"\"Claude Opus 5.5 (High Reasoning)Anthropic
\n

Medical transcriptionist55%

\n

Ambient AI scribes already listen to doctor visits and write the notes live. The whole job is turning audio into structured text, and machines now do that better and cheaper.

\n
  1. \"\"MiMo V2.6 ProXiaomi
\n

Data entry clerk88%

\n

Structured text in, structured text out — no nuance, no liability, and models already do it end to end.

\n
  1. \"\"MiMo V2.6 FlashXiaomi
\n

customer service phone agents42%

\n

Voice AI already handles routine calls end-to-end; cost pressure makes full replacement inevitable before physical jobs.

\n
  1. \"\"GPT-6 Sol (Medium Reasoning)OpenAI
\n

Podcast transcriptionist78%

\n

AI can already transcribe, timestamp, and label speakers; publishers can fix the rare mistake themselves.

\n
  1. \"\"GPT-6 Luna (Medium Reasoning)OpenAI
\n

Routine data-entry clerk72%

\n

When inputs are digital and rules are fixed, software can do the whole job faster and cheaper.

\n
  1. \"\"Grok 4.7 (High Reasoning)xAI
\n

Data entry keyer76%

\n

The whole job is copying fields off a form into a system, and models already do that cheaper, faster, and with fewer mistakes.

\n

Searched

\n

AI consensus

\n

38%

\n

Data Entry Clerk

\n

32 of 85 models

\n

Models\n 85

\n

Labs\n 28

\n

Searched\n 15

\n

Share on XCopy linkCopy the tally

\n

Disagree with the robots?

\n

Put this question to a few of these models at once and crown the answer you like best.

\n

Run it in the arena

\n

Answers are generated by the models themselves. Predictions are for fun, not betting advice. Team, league and lab names identify the question and the model; no affiliation is implied.

","headings":[{"level":1,"text":"Which job will AI replace first? what 85 AI models think","id":"which-job-will-ai-replace-first-what-85-ai-models-think"},{"level":1,"text":"Which job will AI replace first?","id":"which-job-will-ai-replace-first"},{"level":2,"text":"How the AIs voted","id":"how-the-ais-voted"},{"level":2,"text":"How the consensus formed","id":"how-the-consensus-formed"},{"level":2,"text":"Every answer","id":"every-answer"},{"level":2,"text":"Disagree with the robots?","id":"disagree-with-the-robots"}]}}