Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

←

Публикация

Предыдущая публикация
Carl T. Bergstrom

3. I don’t know what the endgame is from these messages. I don’t know whether someone is simply trying to waste my time or whether, once I reply, it leads to some other manner of scam. But spamming seemingly informed and good faith questions of scientists serves as a DDOS attack on the field.

324927
Carl T. Bergstrom

4. As Jevin West and I have been saying for several years now, because a seemingly credible message no longer implies a human sender, LLMs will push us toward more emphasis on reputation and status — and as such serve as an anti-egalitarian force in all aspects of society, science included.

Обсуждение

Ответы

Layer8

True but this is just an intensification of a previously existing trend. I stopped answering my phone to any caller I didn't recognise or responding to private messages from strangers years ago. The number of fraudulent human agents was already an issue before AI ones came into the mix.

Carl T. Bergstrom

Update: Today, a nuanced question about my work on misinformation dynamics. I took twenty minutes to answer. The response was almost immediate and asked further questions, but unlike the original email it was more plainly AI. The sender? A gmail address with from a name with zero online footprint.

Ответ для Carl T. Bergstrom
Carl T. Bergstrom

I can't figure whether someone is training an AI, generally causing mischief, or attempting to waste the time of researchers working in certain critical areas (misinformation, climate change, health policy, etc.)

Ответ для Carl T. Bergstrom
privacyfish

The email piece is what I’d separate: a Gmail address with a plausible From name is identity theater, not identity. If agents can draft or send through an existing account, the sender can look human enough to earn a careful reply. Provenance matters more than the address itself.

Lauren Eastwood

Thank you for everything you post about AI, Carl. It’s important because so very much is flying under the radar right now and shifting our “real” world bit by bit so it becomes more difficult to point to the implications and not sound like Chicken Little.

PlantEvolution 🌱🌾

1/3 I also have the habit of answering every email. I send a short note of acknowledgment even if, e.g., a PhD inquiry is LLM generated.

Ответ для PlantEvolution 🌱🌾
PlantEvolution 🌱🌾

2/3 But I recently had a clearly LLM generated inquiry from a colleague about one of our papers. I ended up having an LLM generate an answer, which I only lightly edited.