Unsurprisingly, Perplexity does not understand consent. Perplexity intentionally ignores robots.txt, and mutating user-agents when blocked by detection systems and (of course) did not reflect that they are not gaining CONSENT from site owners. www.perplexity.ai/hub/blog/age...
I do not say this lightly: Perplexity are aiding fascism with Trump’s propaganda. www.axios.com/2025/08/06/t...
Interesting point of view and raising a lot of questions actually. But couldn't we raise the same concerns with all AI tools ? (OpenAI, Anthropic, Google ?)
For sure — afaik all of the major models are trained on stolen content However a distinction in this case: Perplexity ignore/do not read robots.txt, a common mechanism site owners use to set bot rules. When security systems detect & block, Perplexity mutates user-agent to hide & continue
One could say that at least Perplexity was one of the first to explicitly source the data and sites being parsed. Though I am not really aware of those mechanisms applied to Ai Agents. I will investigate a bit deeper. Thanks.