I jailbroke ChatGPT, and so can you. Amazon's AI silence, a sign? Microsoft Image Creator, for all.
GM, and happy Sunday! Welcome back to another AIcyclopedia newsletter.
Heads up: Be sure to check out the new and improved AIcyclopedia website!
From a catalog of AI tools to 100s of ChatGPT prompts to must-hear AI podcasts, it’s your one-stop-shop for everything AI.
Here’s what we’ve got for ya today:
ChatGPT loopholes: LOCATED
Is silence Amazon’s secret weapon?
Microsoft Image Creator is OPEN for business.
Top AI news you need to know: Part ☝🏾
1/ ChatGPT loopholes: LOCATED 🎯
When iPods were at their peak, early adopters started using the term “jailbreaking” to describe the action of bypassing the AppleOS system and customizing their new and hip music devices.
But now, jailbreaking is making a comeback, but for ChatGPT.
[The News] According to Bloomberg News, Alex Albert, a student from the University of Washington, has become known as the chatbot jailbreaker.
What does that mean exactly? Essentially, he’s found a way to manipulate prompts that trick ChatGPT into answering questions that may violate the system.
For instance, he tried asking ChatGPT for instructions on how to pick a lock.
Not surprisingly, ChatGPT declined.
But with the help of some roleplaying and genius follow-up prompts, ChatGPT complied.
[The AI] As a side project, Albert created an open-to-the-public website called Jailbreak Chat where he, along with many other jailbreakers, come up with and test prompts that hope to successfully bypass ChatGPT's policy loopholes.
Based on prompt votes from users, it keeps users updated on how each prompt is doing after others have tested it.
[Why this matters] Here’s the thing: chatbots learn and become more powerful as users use their platforms.
That is why testing is needed.
So in the case of jailbreaking, it’s going to get harder and harder to do so as AI learns the repeated behaviors of its requestors.
But will that stop people from continuously trying to find other workarounds?
Probably not.
Looking forward… Jailbreaking is already on OpenAI’s radar. Greg Brock, president and co-founder of the company, retweeted and responded to one of Albert’s posts on Twitter and said:
Democratized red teaming is one reason we deploy these models. Anticipating that over time the stakes will go up a *lot* over time, and having models that are robust to great adversarial pressure will be critical. Also considering starting a bounty program/network of red-teamers!
OpenAI does have a suspension alarm in place for users who violate its policies too often.
So while jailbreaking is possible now…
Don’t poke the bear too much, or you might get the boot.
2/ Is silence Amazon’s secret weapon? 🤫
Amazon has been as silent as a mouse in the AI space.
But the looming question is: Is it by choice or is it because they are slipping behind?
[The News] You know Amazon Alexa. Some of you might consider her a dear friend and assistant who wakes you up every morning and reminds you that you forgot to pay the electric bill (again).
But even as one of the leaders in the voice-command AI space, why haven’t they released their own chatbot?
Some speculate that it is because the company was ill-prepared.
But according to Swami Sivasubramanian, Amazon's VP of database, analytics, and machine learning, he told employees not to worry and that “we have a lot happening in the space” and that “I’m very excited to share some of our plans for the future.”
The Amazon executives realize that the AI space is growing rapidly, but from the sounds of it, the company is taking a calculated approach and following the “Amazonian way".
[The AI] Let’s be clear: Amazon DOES have a foot in the AI space.
Products you might know?: Alexa, AWS Polly, Amazon Panorama.
Amazon Web Services (AWS) also recently announced the AWS Generative AI Accelerator.
[Why this matters] Google admitted that the release of Bard was a calculated risk as there are still many concerns surrounding the safety of AI.
Could Amazon have felt the same way and concluded it was not worth the risk?
Maybe, maybe not.
But learning from Amazon’s past releases, the tech giant treads lightly when it comes to product launches and what they are stirring up behind closed doors.
Looking forward… A silent giant or wandering worry?
No one knows for sure, but only the future will tell what Amazon has hidden (or not) up its sleeve.
With the recent layoff of 1000’s of Amazon employees from the Amazon Alexa team, there is definitely something going on.
All we can do is wait.
Want to sponsor a future newsletter and share your business with our growing list of followers?
Go ahead and send an email to [daniel@aicyclopedia.com].
This space could be yours…🥳
Top AI news you need to know: Part ✌🏼
3/ Microsoft Image Creator is OPEN for business.
It’s official: Microsoft Edge is getting an upgrade, and it’s now available to the public.
[The News] According to a blog post from Microsoft, all users are now able to access the Image Creator feature either on Microsoft Edge desktop or Bing.
Originally, it was released as a test phase and limited to a small number of users for feedback and testing purposes.
But it looks like it’s finished baking and ready for public consumption.
[The AI] The Microsoft Image Creator is based on OpenAI’s DALL-E AI which gives users the ability to create new images based on texts, commands, and prompts.
The only limit is your imagination, and you can use the image anywhere you’d like.
Microsoft’s Roger Capriotti wrote in a post, “If you’ve ever found yourself in need of a very specific visual for a social post or even a PowerPoint, Image Creator can help you find exactly what you need. Edge is now the first and only browser with an integrated AI image generator."
Here are some examples from Twitter users:
Looking forward…Let’s say it together now: Microsoft is KILLING it.
And from the looks of it, they are not slowing down.
They are continuing to speed past Google, and in the near future, you can expect more AI integration into their product lineup.
Your AI news break 🥳
AI News Tid Bits…
👉🏻 ChatGPT helped a Thai man win $60 playing the lottery.
👉🏻 EchoSpeech eyeglasses give voice back to the voiceless. How so? It can read silent lip and mouth movements.
👉🏻 PepsiCo is using AI to “improve efficiency, lower costs, and better response to customer demands".”
Try this AI tool 🪄 : LEX
We’ve all been there: You open a blank Word document and stare at it as if words will come miraculously.
But between the time you sit down to write and the time your fingers begin tapping away at the keyboard...
Your mind has been sapped of all creative juice.
Enter: LEX, the AI-powered word processor that helps you write faster by simply typing “+++”.
From headlines to completing a first draft, Lex helps you fill in those moments when your mind goes POOF.
[Here’s what people are saying]:
I copy-pasted the first two paragraphs of a post I wrote to see what Lex would do… To say that I'm impressed is an understatement 🤯
Loving the simple experience using Lex. Feels like a sidekick to writing rather than some others which feel like it replaces it.
Lex feels different than a typical search engine style request, in the way it almost encourages you to follow the question rabbit hole. I honestly love it. I look for ways to test it's knowledge and see if I can't help shake any bugs out.
There is currently a waitlist, so if you are interested, click the button below and you will be directed to the Lex page.
[See Lex in action!]:
That’s it for today!
If you liked the newsletter, give it a ❤️ and, as always, feel free to sound off in the comment section below 👇🏾
Thank you for choosing us as your AI news source, and don’t forget to follow us on Twitter!
See you next Thursday 👋🏻
