100人の人間にDALL・Eのプロンプトを描いてもらった

BlogLeaderboardsWorkforceProductsResearchCareersContactLoginMenuCloseBack to BlogWe asked 100 humans to draw the DALL·E prompts

Quick: which one of these was generated by a human, and which by an AI?

(Answer: AI, human, human, AI.)

OpenAI’s DALL·E image generation system is wild. Imagine a future where every homemade bedtime story can be turned into an IMAX-quality cartoon…

At Surge AI, our bread and butter is using superintelligent humans to train rich, creative AI. But what happens when those machines become advanced enough to augment human capabilities instead? Where will human artistry and creativity fit in then?

Here's an optimistic comment from a visual designer on Reddit:

Artists, designers, photographers, creative types of all flavors will have important roles in this new AI-image world. I'm a visual designer and I honesty can't wait for the engines to get better, or at least to be able to get access to something like Dall-e-2.

I've said this elsewhere, but AI will make the kind of work I do even more valuable because I'll spend less time on grunt work and more time on higher-value creative tasks such as creative/art direction and concept development. If I need series of icons in a certain style for a project, instead of spending hours or days searching vector stock resources for what I need or creating them myself, I can just describe what I want to the AI and have it generate what I need in seconds. Seconds of work instead of days. That makes me insanely more productive. Instead of one project for a client, I've now churned out a dozen in the same amount of time.

To understand this question better, we paid 100 random Surgers – most with average drawing skills! – to spend 15-30 minutes drawing one of the DALL·E prompts. Here are some of our favorites and how they compare.

An astronaut, playing basketball with cats in space, as a children’s book illustration

Teddy bears, mixing sparkling chemicals as mad scientists, as a 1990s Saturday morning cartoon

Interestingly, as one Hacker News commenter pointed out, the human-drawn bears are clearly mad scientists out to conquer the world, while the AI bears lack that crazy, evil gleam.

An illustration of a baby daikon radish, in a tutu, walking a dog

Got a creative task you’d like hundreds of Surgers to work on as well? We believe data labeling isn't about mere sentiment analysis anymore. It's about training AI in the full spectrum of capabilities we want it to attain, whether it's understanding the subtleties of social media, solving math problems, teaching machines human values, learning to program, evaluating search quality, or becoming the next van Gogh. Follow us on Twitter at @HelloSurgeAI to learn more!

Follow us onLinkedinXEnterpriseBench: CoreCraft – Measuring AI Agents in Chaotic, Enterprise RL EnvironmentsHemingway-bench Leaderboard: Because Good Writing Isn't a Checklist of VibesBuilding AdvancedIF: Evolving Instruction Following Beyond IFEval and “Avoid the Letter C”LMArena is a cancer on AIRL Environments and the Hierarchy of Agentic CapabilitiesHow do frontier models perform on real-world finance problems?A Product Take on Sonnet 4.5Is Sonnet 4.5 the best coding model in the world?The Human/AI Frontier: A Conversation with Bogdan GrechukSWE-Bench Failures: When Coding Agents Spiral Into 693 Lines of HallucinationsBenchmarks are brokenUnsexy AI Failures: The PDF That Broke ChatGPTBringing light to the GPT-4o vs. GPT-5 personality controversyDALL·E 3 and Midjourney Fail Astral Codex Ten's Image Generation BetHow Anthropic uses Surge AI to Train and Evaluate ClaudeWe Evaluated ChatGPT vs. Google on 500 Search QueriesAI Red Teams for Adversarial Training: How to Make ChatGPT and LLMs Adversarially RobustHellaSwag or HellaBad? 36% of this popular LLM benchmark contains errorsHow TikTok is Evolving the Next Generation of SearchEvaluating Generative AI: Did Astral Codex Ten Win His Bet on AI Progress?Why Instagram is Losing Gen Z: We Asked 100 Users to Compare TikTok vs. ReelsThe $250K Inverse Scaling Prize and Human-AI AlignmentSearch Behind-the-Scenes: How Neeva Uses Human Evaluation to Measure Search QualityHuman Evaluation of Large Language Models: How Good is Hugging Face’s BLOOM?30% of Google's Emotions Dataset is MislabeledAI Red Teams and Adversarial Data Labeling with Redwood ResearchHumans vs. Gary Marcus vs. Slate Star Codex: When is an AI failure actually a failure?How Surge AI Built OpenAI's GSM8K Dataset of 8,500 Math ProblemsWe asked 100 humans to draw the DALL·E promptsGoogle Search is Falling BehindMoving Beyond Engagement: Optimizing Facebook's Algorithms for Human ValuesHoly $#!t: Are popular toxicity models simply profanity detectors?Is Google Search Deteriorating? Measuring Google's Search Quality in 20225 Examples of the Importance of Context-Sensitivity in Data-Centric AIThe AI Bottleneck: High-Quality, Human-Powered DataFrom the frontier