Retour aux Actualités
EntrepriseBriefing AI Understanding

Les employeurs établissent des règles pour limiter le travail excessif écrit par l'IA

Business Insider rapporte que Polarsteps, Leapsome et Clay ont introduit des directives sur le lieu de travail exigeant que les employés révisent, clarifient et assument la responsabilité de l'écriture assistée par l'IA.

4 min readRead the original reporting
Source-provided image accompanying Employers set rules to curb bloated AI-written work
Rapports attribuésSource enregistrée
Éditeur
businessinsider.com
Lien source
businessinsider.comhttps://www.businessinsider.com/companies-writing-ai-policies-stop-ai-slop-2026-8
Type de source
Reportage d'un média – pas d'un document de première partie.

Ce que nous n'avons pas pu confirmer indépendamment: Cette affirmation est attribuée au point de vente nommé. Nous ne l'avons pas vérifié par rapport à un document de première partie. (businessinsider.com)

ContexteComprenez cela en 60 secondes

Commencez ici

Testez-vousQuiz sur l'éthique de l'IA

Que s'est-il passé

Business Insider reports that several employers are responding to bloated, unclear AI-assisted workplace writing with formal internal policies rather than banning AI. Polarsteps, Leapsome and Clay reportedly require employees to disclose AI use in some contexts, apply critical thinking, prioritize clarity and stand behind every sentence. The source says these policies are internal and their reported effects have not been independently confirmed.

Business Insider reports that Amsterdam-based travel-app company Polarsteps issued principles earlier in 2026 for its roughly 100 employees. The guidance permits brainstorming and drafting with AI but expects employees to disclose AI use in relevant cases and ensure that final copy reflects their own reasoning and voice. The article says consumer-facing app copy is generally produced by professional in-house writers.

Business Insider reports that Leapsome rolled out an AI-writing policy in mid-August for its 130 employees in Berlin and New York. The policy emphasizes clarity, reasonable length and accountability, and says employees should not use “AI wrote it” as an explanation for ambiguous writing.

Business Insider reports that Clay first shared its policy with an engineering team of about 100 people in mid-June and expanded it to all 550 employees in early August. The policy requires employees to stand behind each idea and sentence and make documents worth reading. The article says the policy followed five overtly AI-generated documents sent to engineer Sophie Alpert in one week.

Business Insider cites a late-2025 BetterUp and Stanford Social Media Lab survey in which 40% of 1,150 full-time U.S. desk workers reportedly said they had received AI-generated “slop” from a coworker in the previous month. Respondents reportedly said handling each instance took nearly two hours on average. The survey results were not independently confirmed in the supplied source.

Détails de la source: businessinsider.com ↗

Pourquoi c'est important

The reported policies address a practical workplace problem: AI can reduce drafting effort while increasing the time colleagues spend reading and correcting vague or unnecessarily long documents. The approach also places responsibility on the human employee instead of treating AI use itself as the misconduct. Business Insider’s reporting suggests that employers are beginning to define acceptable AI assistance through standards for judgment, attribution and usefulness, although the evidence comes primarily from company leaders and is not independently verified.

The reported rules distinguish using AI as a drafting aid from passing unexamined output to colleagues as finished work. That distinction is relevant to organizations trying to preserve human accountability while allowing productivity tools.

Longer documents can impose a recurring cost on every reader, while editing for clarity can require more effort from the author once. Business Insider presents this as the reasoning behind Clay’s policy, but the article provides no independent measurement showing that the policy has produced those savings.

The article reports mostly positive reactions from leaders behind the policies, including requests from some Leapsome employees for stricter limits on AI in internal messages and marketing copy. Those reactions are reported by the company representatives and do not establish wider employee sentiment.

Interactive Mechanism

Mécanisme interactif : comment cela fonctionne réellement

Explorez de manière interactive la technologie sous-jacente à ce développement.

Thinking Budget (Test-Time Tokens):1,024 tokens
Complex Accuracy79%Math & Code Logic
Latency3.2sTime to first full output
Inference Cost$0.0092Per query estimated
Reasoning StyleStep VerificationInternal chain depth
Active Thinking Trace:
1Deconstruct user problem into formal constraints
2Propose candidate hypotheses & step-by-step calculation
3Self-correction: Backtrack and refute subtle edge cases
4Exhaustive consistency check & final output synthesis
Core takeaway: Test-time compute fundamentally changes AI economics. Instead of only scaling during pre-training, giving reasoning models more tokens at inference time allows them to systematically solve PhD-level STEM problems.
Vérification de concept interactive+10 Points
AI Ethics Quiz

Why can ethical evaluation not be reduced to one model score?

Que regarder ensuite

Watch whether employers publish measurable standards for AI-assisted writing, disclose how policies are enforced and show whether the rules reduce reading time or improve decision-making. It is also unclear how broadly these policies will extend beyond internal documents, whether employees face penalties for violations, and whether disclosure requirements will apply consistently across teams and communication channels.

The source does not provide public copies of the full Polarsteps, Leapsome or Clay policies, enforcement procedures, violation rates or evidence that the rules changed productivity or document quality.

Business Insider reports positive effects at Polarsteps and positive worker reactions at the other companies, but these claims come from executives or policy authors. Independent employee feedback and before-and-after data remain unknown.

The article does not report access restrictions or pricing because these are internal workplace policies rather than consumer products. It is unknown whether the companies will extend the rules to external communications, marketing material or employee-to-employee messages.

Guides et quiz associés

Éthique de l'IAModèles d'IA expliquésPrompt EngineeringTestez ce que vous savez : essayez un quiz gratuit sur l'IARecherchez un terme d'IA dans notre glossaireSuivez le suivi du financement de l'IA
Vous avez trouvé cela utile ?