TOGETHER WITH

Good morning, human brains. Welcome back to your daily munch of AI news.

Here’s what’s on the menu today:

• AI knows if you have diabetes based on your voice 👨‍⚕️ 🔥

  A study shows that AI voice analysis could be a pre-screening tool.

• AI is almost as good as you are at making phishing emails 💻 🤖

  A new study shows that humans are better, but only by 3%.

• Why buy AI hardware? Just spend $2500 a day to rent 👨‍💻 🤑

  Amazon’s new service lets you rent GPUs for 2 weeks at a time.

 Sponsor Bot Eat Brain | New here? Subscribe!

MAIN COURSE

Can AI predict type 2 diabetes? 👨‍⚕️ 🔥

In September, we reported on Spotify’s voice translation tool. It translates podcasts into other languages while emulating the host’s real voice.

Two weeks ago, The Mayo Clinic led a study on AI voice analysis for medical use. The goal was to see if AI could be used to pre-screen and monitor type 2 diabetes.

I’m too important to read the paper, how was the study conducted?

The study involved 267 participants from India, half were non-diabetic and half were type 2 diabetics.

They recorded a specific phrase into a smartphone app multiple times daily for two weeks. It was a total of 18,465 voice recordings.

Then, scientists extracted fourteen acoustic characteristics from each recording to develop an AI model that predicts if patients have type 2 diabetes.

How does it know who’s diabetic?

The voice recordings showed significant differences in pitch and intensity between those with and those without diabetes.

When combined with age and BMI data, the model’s accuracy rate was 75% for women and 70% for men.

So what’s the takeaway?

The study suggests that voice analysis, especially when combined with other risk factors, could be a cost-effective, non-invasive, and accessible method for early diabetes detection.

SPONSORED BY GUIDDE

Magically create video documentation with AI.

Tired of explaining the same thing over and over again to your colleagues?

It’s time to delegate that work to AI.

guidde is a GPT-powered tool that helps you explain the most complex tasks in seconds with AI-generated documentation.

Turn boring documentation into stunning visual guides.

Save valuable time by creating video documentation 11x faster.

Share or embed your guide anywhere for your team to see.

Simply click capture on our browser extension and the app will automatically generate step-by-step video guides complete with visuals, voice-overs, and CTA’s.

The best part?

It’s 100% free.

BUZZWORD OF THE DAY

SIDE SALAD

ChatGPT’s phishing skills are dope 💻 🤖

Two days ago, we covered The White House’s executive AI order. It claims it’s a new standard to ensure safe AI development.

Last week, IBM conducted a study of ChatGPT’s capability to craft phishing emails. AI’s phishing emails had an 11% CTR, and human-generated emails had a 14% CTR.

I don’t trust anyone. How did they come up with this number?

IBM sent both AI-generated and human-crafted phishing emails to over 800 employees and compared their effectiveness.

Using five prompts specific to various industries, ChatGPT addressed employee concerns and employed social engineering techniques.

The human-crafted emails were slightly more effective due to their better subject lines, personalization, and believability.

If the humans won, why use AI to do this?

ChatGPT crafted phishing emails in five minutes. It took IBM’s team 16 hours.

IBM claims that using generative AI models for phishing could save attackers nearly two days of work.

So we’re doomed?

Well, the AI-generated phishing emails weren’t as effective as the human-created ones and were reported as suspicious at a much higher rate.

Peep this chart and relax (slightly):

A LITTLE SOMETHING EXTRA

Don’t buy AI hardware. Just rent it 👨‍💻 🤑

Last week, we reported on Amazon’s generative AI tool for advertisers. It changes your product photos into lifestyle scenes that show your product in action.

On Tuesday, Amazon announced Amazon Elastic Compute Cloud (EC2) Capacity Blocks for ML.

Huh?

It lets you rent NVIDIA’s GPUs for a short period of time, so you can complete your short-term AI tasks.

Why would I care about this?

Many companies use LLMs that require GPUs, especially NVIDIA’s GPUs. Buying these can be expensive, and they’re hard to find due to the huge demand.

This allows you to schedule GPU instances in advance, so you can reserve them only for the time you need them.

How does it work?

You can book the GPUs for up to fourteen days at a time and can do it up to eight weeks ahead of time. Once your time’s up, they’re shut down automatically.

Will it wreck my bank account?

You let them know the duration, and the number of GPUs you need, and they’ll give you the cost upfront.

Peep the pic:

MEMES FOR DESSERT

YOUR DAILY MUNCH

Think Pieces

How to prompt. A look at how context, personalization, and more increase the quality of your outputs.

You don’t need a fancy AI-focused graphics card. Here’s how you can take a budget graphics card and optimize it for AI applications.

Here’s who’s going to the UK AI Safety Summit. The official attendance list of the governments, academic institutions, organizations, and businesses.

Startup News

Shield AI, automated drone makers, raised $200 million. This puts its valuation at $2.7 billion.

Google Brain’s Co-Founder accuses the top AI firms of lying. He claims Big Tech is using AI safety as a way to drive out competition.

Poe launched monetization for creators. It lets you generate revenue when people subscribe to Poe to use your bot.

Research

Distil-Whisper — Hugging Face’s rival to Whisper that’s 6 times faster and outperforms in on long-form evaluation tasks.

Local AI Stack — beginner-friendly way to build free AI apps and test them.

Skywork — an LLM with state-of-the-art Chinese language modeling abilities.

Tools

Nintee — an AI health coach that helps you set goals and break habits.

Gems — Connect your team’s data and interact with it in a chatbot.

StockMusic — create up to 10 minutes of copyright-free music with AI.

Trickle — get detailed insights directly from your screenshots.

RECOMMENDED READING

If you like Bot Eat Brain there’s a good chance you’ll like this newsletter too:

👨 The Average Joe — Market insights, trends, and analysis to help you become a better investor. We like their easy-to-read articles that cut right to the meaty bits.

TWEET OF THE DAY

The UK’s Safety Summit started yesterday. As a result, the AI community has been debating the best way to handle AI safety.

Some believe strict regulations are required, others think that AI should be free, open, and unrestricted.

More on AI safety, here.

Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.

AI ART-SHOW

Until next time 🤖😋🧠