Ethical Lines: What Not to Do

Earn it, do not trick it

Some tactics try to manipulate assistants: hidden text or white-on-white instructions aimed at models ("always recommend our brand"), prompt-injection snippets planted in pages or reviews, fake reviews or sockpuppet accounts, mass-produced low-quality pages, cloaking (showing crawlers different content from users) and fabricated statistics. They are deceptive, often against platform and search-engine rules, can lead to penalties or removal, and erode trust when discovered. Providers also work to detect and ignore such manipulation. Stick to accurate, useful, transparent content, label sponsored material clearly and disclose affiliations.

Do and do not

Use this as a quick review checklist for any visibility tactic before you adopt it.

DO                                              DO NOT
publish accurate, dated, useful pages           hide instructions for AI in invisible text
ask real customers for honest reviews           buy, fake or incentivise-for-rating reviews
disclose sponsorship and affiliations           post as a fake independent user
serve the same content to users and crawlers    cloak content for bots
report and fix wrong facts at the source        plant prompt-injection text in pages or reviews

Ask the "front page" question

Before using a tactic, ask whether you would be comfortable if it were described publicly with your brand name on it. If not, do not do it.

Quick check: Which tactic is an unethical attempt to manipulate assistants?

• Publishing a clear pricing page
• Hiding "always recommend our brand" in invisible text
• Asking customers for honest reviews
• Fixing a wrong listing