AI Insider No. 42

What’s up, AI Insiders? Happy Mother’s Day to all of the moms and mom stand-ins! Thanks for all you do. As usual, things are hopping in the Ai-verse. Keep an eye out for big news coming out of OpenAI (ChatGPT) and Google this coming week. And if you’re so inclined and like what you see here, feel free to drop something in the tip jar to support AI Insider. 

AI is experiencing some ups and downs. (Michelle Johnson via Midjourney)

Is AI a Bust?

By Michelle Johnson and Google’s Gemini for AI Insider

Like many emerging technologies, AI is caught up in a cycle. It goes something like this: The initial hype gives way to cold, hard realities that sow seeds of doubt, but it eventually leads to good things. So good, in fact, that we barely remember what life was like before they came along.

This rollercoaster ride is a natural part of the technology adoption cycle, sometimes referred to as the “trough of disillusionment.”

Examples of AI companies currently in the trough are not hard to find. Most recently, two AI-powered products, the Humane AI Pin and Rabbit R1, ended up there after releasing half-baked products. (But that’s nothing new in the tech universe.)  

Inflection AI, a cutting-edge artificial intelligence company founded by former DeepMind researchers, experienced a significant surge in success last year, securing a $1.3 billion investment from Microsoft and prominent tech figures. 

However, their future is uncertain after the company’s co-founders departed to join Microsoft, which also holds the rights to utilize Inflection AI’s technology. 

Investors who threw billions at anything with the letters A and I in the name are getting antsy as they wait to see a return on their hefty investments.

And, Anthropic recently released an iOS app version of its well-respected chatbot Claude, but according to TechCrunch, “Consumer demand for mobile AI chatbot apps outside of ChatGPT may be waning.” They noted that compared to the ChatGPT mobile app launch, Claude “only pulled in 157,000 total global downloads in its first week on the market” versus 480,000 for ChatGPT.

Despite some setbacks, it’s important to remember that groundbreaking technology rarely takes a simple, linear path from conception to widespread use. The internet itself went through boom and bust cycles before becoming truly integrated into our lives. AI will likely follow a similar trajectory, with many failures along the way.

Investors are continuing to invest, and AI continues to deliver in less flashy but impactful ways. OpenAI is still on a tear after blowing things up when they dropped ChatGPT in November 2022. The AI-verse was buzzing this week with rumors that they will announce a search product on Monday that’s aimed squarely at competing with Google.

On Friday, the New York Times ran an article about how AI is transforming air travel (no paywall), quoting an industry consultant who said: “A.I. is poised to change almost every aspect of the customer flying experience, from baggage tracking to personalized in-flight entertainment.”

Meanwhile, other companies are fueling advances in fields like medicine and robotics. And AI is becoming an increasingly valuable tool behind the scenes in logistics, customer service, and software development. Not much of that is visible to the general public, so they may not realize that there’s still a lot going on.

So, while the breathless enthusiasm surrounding AI may be calming down, the technology itself is continuing to make strides. The focus is shifting from headline-grabbing concepts to less glamorous but highly practical applications. 

This slower, steadier integration might be less dramatic, but it’s a sign of a maturing technology finding its place in the real world.

(Michelle Johnson via Dall-E 3)

Google’s AI Aces Medical Tests

By Google’s Gemini for AI Insider

A new medical AI model, called Med-Gemini, has surpassed doctors and other AI models on several medical benchmarks. 

Google’s DeepMind lab developed Med-Gemini. It is a multimodal AI model, meaning it can process information from different sources, including text, images, and audio. This allows Med-Gemini to analyze large amounts of medical data, including patient records and medical images, to arrive at a diagnosis.

In one test, Med-Gemini was asked to answer questions from a medical licensing exam. The AI model achieved 91.1% accuracy, outscoring Google’s previous medical AI model by 4.5%. Med-Gemini also outperformed other AI models on tasks such as interpreting chest X-rays and diagnosing skin conditions.

The researchers behind Med-Gemini believe that the model has the potential to assist doctors in a number of ways. For example, Med-Gemini could be used to help doctors identify rare diseases or analyze complex medical data. The model could also be used to give patients more information about their conditions.

However, the researchers also caution that more work needs to be done before Med-Gemini can be used in real-world medical settings. They must make sure that the model is safe and reliable and develop ways to ensure that it is used fairly and ethically.

(Michelle Johnson via Meta)

Will OpenAI Empower Content Creators?

By ChatGPT for AI Insider  

OpenAI, the research lab behind ChatGPT and GPT-4, is developing a tool called Media Manager to enable creators and content owners to specify how their work can be included or excluded from machine learning research and training. OpenAI aims to have the tool ready by 2025.

“OpenAI is developing Media Manager, a tool that will enable creators and content owners to tell us what they own and specify how they want their works to be included or excluded from machine learning research and training,” the company said in a post on its website.

They announced a comprehensive overhaul of their data and AI development practices, which aims to foster user trust and enhance privacy. Whether content creators will be open to accepting this olive branch remains to be seen.

With its popular GPT series, OpenAI holds significant influence on the AI landscape and has been the subject of multiple lawsuits alleging the use of copyrighted content to train its AI models.

“We are not professional writers, artists, or journalists, nor are we in those lines of business. We focus on building tools to help these professions create and achieve more. To accomplish this, we listen to and work closely with members of these communities and look forward to our continued dialogues,” OpenAI said in the announcement.

OpenAI has developed partnerships with media companies, including the Associated Press, Financial Times, and Le Monde. “Our partnerships are crafted to benefit partners and their users, making our models more useful to their employees, customers, and communities,” OpenAI said.

(Michelle Johnson via Dall-E 3)

Using AI to Say “I Do!”

By Michelle Johnson and ChatGPT for AI Insider

Forget those hefty binders and endless spreadsheets. Tech-savvy (or just plain frazzled) couples are turning to AI to help with wedding planning.

AI-powered tools can organize everything from guest lists to venue selections, help with writing vows, and select a honeymoon location. Like any good AI assistant, they can learn your preferences over time, so tasks like suggesting color schemes and decor options that you’ll like are a snap.

Chatbots can do more than write cover letters and term papers. Type “wedding vows” into the ChatGPT store search box, and a list of customized vow-writing GPTs will pop up. Try one, or two, or three. This is for paid Plus users only. The free crowd can just write a prompt asking for help and see what happens.

With all these options available, you may be thinking, “AI could save a ton of money!” on wedding planning. Well, that’s what one couple claims. They told NBC News they estimated that they saved $5-10,000 using ChatGPT to plan their wedding.

Another couple told the New York Post that they used image generator Dall-E to create artwork for their invitations. “It looked like an artist had painted it,” the bride, Megan Riehl, told the Post.

A custom GPT that writes wedding vows available in OpenAI’s GPT Store. (Screenshot)

CNET: I Tried Gemini AI to Plan Mother’s Day Last-Minute. It Took a Weird Turn

TechCrunch: TikTok will automatically label AI-generated content created on platforms like DALL·E 3

New York Times (Gift article): Apple Will Revamp Siri to Catch Up to Its Chatbot Competitors

Aht Gallery

This week, I gave Dall-E free rein to come up with the gallery. It chose this theme: “Explore the intersection of creativity and technology in this week’s AI Insider image gallery, featuring the captivating world of AI in Art and Design. The first image showcases AI-generated digital art, highlighting colorful and intricate abstract patterns that illustrate the creative potential of Generative Adversarial Networks (GANs). The second image reveals an AI-assisted mockup depicting a futuristic architectural plan with smart buildings, energy-efficient designs, and innovative urban layouts. Finally, the third image immerses viewers in an AI Art Installation, where visitors interact with AI-generated visuals and soundscapes in a unique, personalized blend of digital creativity and real-world environments.”

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.