Perplexity, the rising star in the AI search field, is rapidly escalating its conflict with e-commerce giant Amazon! The two companies confirmed on Tuesday that Amazon has officially "expelled" Perplexity's AI shopping assistant Comet, demanding it be removed from Amazon's online store. Amazon's anger stems from repeated warnings that went unheeded: Comet was shopping in the marketplace like a "ghost," refusing to reveal its AI agent's true identity, openly violating Amazon's terms of service. After reaching the limit of its patience, Amazon finally sent a strongly worded "cease and desist" legal letter.

Perplexity quickly responded fiercely, publishing a blog post titled "Bullying Is Not Innovation." The article complained, "This week, Perplexity received aggressive legal threats from Amazon, demanding us to prohibit Comet users from using its AI assistant on Amazon. This is Amazon's first legal attack on an AI company and a threat to all internet users!"

The core of this debate lies in a philosophical question: Who exactly is an AI agent? Perplexity insists that since its agents act according to human users' instructions, they should automatically have the same permissions as human users, meaning they don't need to "reveal their identity." However, Amazon responded sharply, pointing out that other third-party agents representing human users (such as food delivery apps or online travel agencies) clearly state their identities.

If, as Amazon says, Perplexity only needs to simply reveal its agent's identity, the issue could be easily resolved. But the "devil is in the details." Perplexity fears that once Comet removes its "disguise," Amazon, which has its own shopping robot Rufus, may immediately shut it down completely. Amazon's statement also carries a threat: "We believe that third-party applications that represent customers in purchasing goods from other companies should operate transparently and respect the service provider's right to decide whether to allow them to participate."

Perplexity revealed what it believes is Amazon's real motivation: AI robots are the "enemy" of ads and product placements. An AI assistant assigned the task of buying a laundry basket would not get distracted like a human, nor would it be tempted to buy more expensive upgrades, nor would it be "accidentally" persuaded to purchase a bestseller or a discounted pair of headphones while browsing.

In fact, this kind of "familiar" conflict occurred several months ago. At that time, network infrastructure provider Cloudflare released a study accusing Perplexity of deliberately "masking" its identity to bypass website AI robot blocking requests (bots.txt) for content scraping. Interestingly, this time many people came to Perplexity's defense, arguing that the AI was simply following user instructions (visiting a specific website when asked), which is no different from a human browser.

However, Perplexity's "suspicious" method of hiding its identity at that time also created hidden risks. As the media pointed out, that incident was just a warning, indicating the chaos we might face if the "AI agent world" predicted by Silicon Valley actually arrives. When consumers and businesses fully outsource tasks such as shopping, travel bookings, and restaurant reservations to AI assistants, should websites completely block them, or should they find a new way to coexist?