As AI slop floods the internet, it’s only a matter of time before more of it starts seeping into everyone’s inboxes. Pangram now wants to make those AI-generated emails easier to spot.
The popular AI-detection platform, known for flagging AI-generated text and images, launched a new integration Wednesday that can scan incoming Gmail messages and label them based on how much of it appears to have been written by AI.
Once users connect their Gmail account, Pangram checks new messages and labels any flagged email as “AI” or “Mixed” by default. Users can also turn on labels for human-written emails.
“New messages are checked and labeled as they arrive. Pausing the integration will stop new scans. Existing labels stay in Gmail,” the company wrote on its website.
Users can exclude specific senders from being scanned. A dashboard also tracks how Pangram labeled incoming mail over the last 24 hours, seven days, or 30 days.
“I’ve been using Pangram’s Gmail inbox labeler for about two months now. It’s been pretty huge to be able to see cold emails labeled as AI and use this info to inform how I want to engage,” Pangram co-founder and CEO Max Spero wrote Wednesday in a post on X.
Spero said there was also significant internal discussion about privacy because of the sensitive nature of email data. The company ultimately landed on zero data retention policy, meaning the company does not retain any email data after processing it or use it for training.
Additionally, Pangram’s website says its detector does not receive information about an email’s sender, recipient, subject line, or other metadata.
Spero also said it was important to give users control over what happens after an email gets flagged. His current personal preference is for the tool to simply label AI-generated emails while leaving them in his inbox, though he said that could change in the future “if AI emails ramp up.”
As a result, users can choose to automatically archive emails labeled as AI, or just send them straight to spam.
However, that comes with some risk. AI detectors are getting better, but false positives are still possible.
Pangram claims its detector can identify AI-generated text with 99.98% accuracy, and an independent study that compared multiple AI-detection tools found Pangram to be the most effective.
Still, Spero recently told Gizmodo that AI-detection tools will always have to play catch-up with increasingly capable generative AI models.
“I don’t think we’ll ever get to one hundred percent accuracy,” said Spero. Instead, the all the company can do is “accumulate more certainty using more data.”
Read the full article here
