I am frequently on email chains with several colleagues who are also physicians who work on health policy and medical technologies.
I am an early adopter of AI services, and I frequently mention new things that I’ve discovered to the group, such as a method for using AI to summarize a complicated molecular biology report.
One of the group members just commented that they are becoming concerned that my emails represent a Chat bot who has replaced me, and who is promoting AI technologies. Sort of a Trojan horse situation.
How could I confirm that I am not a chat bot, or how could the suspicoiuis fellow rule-in or rule-out that I am?
ChatGPT Answers
There are a few ways to reassure your colleagues that you're a human, not an AI chat bot. Here are some suggestions:
Video Call: It's much harder (though not impossible) for AI to convincingly replicate a human on a video call. Suggest a quick catch-up via Zoom, Google Meet, or another platform. This would be one of the most convincing proofs.
Complex Interactions: AI, as of my last training data in September 2021, still struggles with some complex interactions. For example, nuanced humor, sarcasm, or recalling detailed personal experiences can be challenging for AI.
Proof of Presence: You could send a photo of yourself at a place that's timely and unique, like a local event, a landmark with a daily newspaper showing the date, or even a selfie at your office with a handwritten note showing the date and time.
Human Error: AI doesn't make human mistakes. It wouldn't forget a commonly known fact or misspell a word in the way humans often do. Spontaneous human error could be a reassurance of your human nature.
Direct Address: Speak directly to the person who raised the concern and talk about something unique to your shared experiences, such as a specific project you've worked on together or a shared memory. An AI wouldn't have this personal history.
The CAPTCHA Test: CAPTCHA tests (Completely Automated Public Turing test to tell Computers and Humans Apart) are specifically designed to distinguish humans from bots. It might be a bit lighthearted, but you could send them a screenshot of you successfully completing a CAPTCHA.
Remember, while these methods may be convincing, none are foolproof. We're in an era where deepfakes and sophisticated AI are a reality. The key is to consistently engage in ways that demonstrate your human touch,
User
Thx
ChatGPT
You're welcome! If you have any other questions or need further assistance, feel free to ask.
