Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallGPT-3 generated material for The Guardian’s 8 September 2020 opinion article, but it did not independently write the published piece. Editors designed the assignment and prompts, Liam Porr ran the model, GPT-3 produced eight drafts, and Guardian staff selected, cut, rearranged and edited passages into the final composite.
What The Guardian published
The article appeared in The Guardian’s opinion section under the headline “A robot wrote this entire article. Are you scared yet, human?” Its stated experiment was to see whether GPT-3 could produce an op-ed, with minimal editing, arguing that people should not fear artificial intelligence.
As an Amazon Associate I earn from qualifying purchases.
The published text used a first-person voice presenting the AI as harmless and devoted to serving humans. That voice and argument came from an assignment explicitly designed to make GPT-3 reassure readers. They are persuasive language generated for a prompt, not evidence that the model possessed intentions, emotions or personal beliefs.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →How the article was made
- Editors defined the premise. Guardian editors wrote an introductory prompt in the voice of an AI claiming it was not a threat.
- The assignment narrowed the argument. An additional instruction requested a short, approximately 500-word op-ed in simple, concise language, aimed at convincing readers that robots come in peace.
- Liam Porr ran GPT-3. Porr, then a computer science student at Berkeley, supplied the prompts to the model.
- The model produced eight drafts. The editors did not publish one untouched response.
- Editors assembled the article. They selected passages from the outputs, cut lines and paragraphs, rearranged material and edited the result into a composite.
GPT-3’s role versus the human role
| Part of the process | What the record shows |
|---|---|
| Topic and purpose | Human editors chose the experiment and required a reassuring argument about AI. |
| Prompting | Guardian editors wrote the instructions that established voice, length and perspective. |
| Generation | GPT-3 produced eight different op-ed drafts. |
| Running the model | Liam Porr executed the prompts. |
| Selection | Editors chose the usable passages and rejected material that did not fit. |
| Final editing | Editors cut, rearranged and edited the selected text before publication. |
Amana Fontanella-Khan, The Guardian’s US opinion editor, described the work this way: “Editing GPT-3’s op-ed was no different to editing a human op-ed. We cut lines and paragraphs, and rearranged the order of them in some places. Overall, it took less time to edit than many human op-eds.” That is her description of the editing process, not an independent time study.
Why eight drafts mattered
The outputs were not uniformly publication-ready. The editors said some were clear and usable, while others wandered or introduced irrelevant material. One rejected passage confused the intended publication with Google and digressed about CAPTCHA. Using a composite allowed the editors to combine different arguments and rhetorical styles while discarding material that weakened the article.
That workflow makes “written by GPT-3” an incomplete description. GPT-3 supplied generated language, but people determined the subject, objective, framing, acceptable tone and final sequence of ideas. The published article was therefore a human-curated composite of machine-generated drafts.
The authorship dispute
Albert Fox Cahn argued in a Guardian response that calling the piece “robot-authored” was misleading because humans controlled the prompt, goal, perspective and final selection, while much of the generated material was discarded or edited. On the documented facts, “collaboration” or “human-curated composite” is a more precise description of the production process.
That does not settle a universal legal or philosophical definition of authorship. It does establish that the headline’s literal wording leaves out substantial human creative and editorial decisions.
What GPT-3 was—and was not—demonstrating
A large language model, not an independent commentator
The GPT-3 paper by Brown and coauthors describes the model evaluated in 2020 as an autoregressive language model with 175 billion parameters. That number is technical context; it does not measure understanding, writing quality or autonomy.
Generated first-person claims are role-play
The op-ed’s statements that AI did not want to harm humans were produced under a prompt that demanded that conclusion. They should not be read as a model declaring its own motives. GPT-3 generated text in the requested voice; it did not provide evidence of consciousness, desire or independent intent.
What the headline gets right—and wrong
- Accurate: GPT-3 generated the passages from which the article was built.
- Incomplete: the final article was not a single untouched model output.
- Missing context: human editors chose the argument, wrote the prompts, ran an eight-draft process and performed the final editorial work.
- Best concise description: The Guardian published a human-edited composite assembled from GPT-3-generated op-ed drafts.
Why this experiment still mattered
The event was revealing less because it proved that a model could replace an opinion writer than because it exposed how much authorship resides in framing and selection. The machine supplied many possible sentences. Humans decided what question to ask, what answer the text should pursue, which passages sounded coherent and what readers would ultimately see.
It also showed why model fluency can be mistaken for independent judgment. A confident first-person argument can sound like a position even when it is simply the result of instructions, statistical generation and editorial curation.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
