When I first started using ChatGPT, I couldn't stop talking about it and sharing it with everyone I knew. But sometimes, I'd be hanging out with friends and just want to quickly ask ChatGPT a question over drinks. That's when I thought: what if this was all on my WhatsApp? What if I could message it anytime, anywhere? – Bertha Kgokong
Bertha then went and created MobileGPT – a WhatsApp integration that utilizes both the GPT and DALL-E APIs from OpenAI. Thus, granting us the convenience of accessing generative AI directly from the world’s most used messaging app.
[WhatsApp has over 2.2 billion users who spend an average of 37 minutes a day on the app and exchange over 100 billion messages daily.]
Honestly, it was only a matter of time before the LLMs entered the world of text messaging. Furthermore, I predict that someone will soon build a similar product linked to a 1-800 phone number – charging people a few pennies per message – but giving you the ability to text with an LLM from any messenger.
The convenience factor is excellent. I think that MobileGPT proves how LLMs will be integrated into almost every piece of software we use. However, I worry about AI invading our regular day-to-day communication with friends and family.
Honestly, is it really a good idea to automate our texts?
A few months back, I was texting with Cowdrey on some ideas for an upcoming book and he asked me a really difficult question. I was tired and didn’t want to think too hard, so I had ChatGPT generate a response for me.
He immediately knew. The tone was way off and used words I never use. The formatting looked nothing like my texts because I never send long paragraphs. And it had no real personality or visionary take that I usually deliver with.
I got caught red-handed and he replied, “If you ever respond to me with ChatGPT again, I’m blocking your number.”
Harsh? Yes. But I completely understand where he’s coming from. And I actually think he’s right that we should fight against AI coming into our interpersonal communication.
Imagine spilling your heart out to someone only to find out you were texting with AI the entire time. Imagine catching up with a friend only to realize that an AI was filling in the details of their life updates. You’d be devastated.
It’s one thing to build a friendship with AI (like Replika) when you know you’re communicating with an AI. But being deceived is not acceptable. Who wants to text an emotionless being that’s faking their feelings?
On the other hand, I wonder if Apple Business Chat would be a good use case for bringing GPT into messaging. Business Chat gives you the ability to iMessage directly with companies. And this could provide a way for these chats to feel less like you’re talking with an auto-response bot.
Not to mention, every iPhone has an ML chip for on-device machine learning processing, and the newer iPhones have several of these chips. They will be able to run Chat or Stability AI right from the device, at no extra cost, and with no real issues with latency.
Apple needs to make a play in generative AI, and as always, they’re really well positioned to succeed.