Hey everyone!

Happy Thursday! Welcome back to a bunch of incredible AI updates that have happened in the past few days!

Here’s what we’ll learn about today:

• Open AI’s CEO defends AI in Congress hearing🏛️

• Google’s new AI update will transform your search experience

• Microsoft says their new AI shows signs of Human Reasoning

• The ChatGPT App!

Reading time: 4 minutes 58 seconds

Enjoy!

News Reel:

Highlight 1: Open AI’s CEO testifies in Congress hearing about AI safety.

Big news from Congress! They had OpenAI's CEO, Sam Altman, show up for the first time ever to face Congress 🏛️. Senators got a chance to grill him about OpenAI's take on regulation. According to Senator Dick Durbin from Illinois, it was a pretty historic session. The main focus was on the potential dangers of AI models like ChatGPT and how lawmakers should handle them 🧑🏽‍⚖️.

Altman's comments surprised some senators. Here are the main points:

• He suggested creating a federal agency to oversee and control the development of AI 💻.

• He also believed that creators should be paid when their work is used to train AI systems 💸.

• He agreed that people harmed by AI products should have the right to sue the developers 💥.

When asked about the "AI pause" letter we mentioned some newsletters ago, which called for a six-month break before deploying more powerful systems than GPT-4, Altman didn't seem too concerned… because OpenAI spent more than six months evaluating GPT-4 before using it and had no plans to release another model in the next six months 😂.

During the hearing, Altman and other speakers mostly agreed on various issues, like:

• They supported privacy protection 🪧.

• Stricter government oversight, and third-party audits 🏛️.

• the United States should have a privacy law similar to Europe's 🌎.

• Altman disagreed with the idea that consumers should be able to opt out of having their publicly available data used in training data sets ❌.

• Altman didn't firmly state that OpenAI opposed offering an ad-supported version of its GPT products (like ChatGPT) 📢.

Exciting stuff happening in the AI world, and it's great to see the CEO of a huge company taking the hot seat and sharing his views with the public. The Twitter thread above has detailed information with the highlights of the hearing, enjoy! And stay tuned for future updates on this topic! 😀

Highlight 2: Google’s new AI update will transform your search experienceRobot solving puzzles

Google is focusing on the transformative potential of AI across its platforms, as highlighted at the recent Google I/O event. 📈 One of these focuses was to make Google search more conversational and interactive 🗣️, making it more in-depth and providing more relevant search results. Here’s a summary of the new features:

• Gmail and other first-party Workspace apps (apps that Google owns) users will have a more detailed email composition due to the integration with Google’s Bard, which could facilitate tasks like flight cancellations and refunds 🤖.

• Conversational Mode 🎙️ in Google Search allows users to search as if having a conversation with a bot by considering the context of previous inquiries. After a search, Google will provide a list of follow-up questions that enable you to delve deeper into specific topics. It takes into account the context of your previous inquiry and removes the necessity to rephrase it.

• In-depth product results in Google Search extend this conversational mode to offer recommendations based on specific features, with links to products, reviews, and pricing information.

• In-depth search results leverage generative AI to provide detailed answers to complex queries, including comprehensive answers, relevant links, and the sources utilized.

• Google also offers content generation upon request, helping users create names, quizzes, social media posts, poems, etc. 👩🏼‍🎨

• Finally, users can request personalized fitness plans, and receive weekly instructions and advice on reaching goals, with additional informational videos in search results. 🏋🏼

Google is widening the gap between itself and Microsoft. Which side are you on?

Here’s the detailed article for everyone interested.

Highlight 3: Microsoft says their new AI shows signs of Human Reasoning

It’s getting real… Microsoft researchers conducted experiments with an AI system, which provided surprising answers to a puzzle requiring an intuitive understanding of the physical world among other several features 🧠. Here’s what happened:

• The researchers published a research paper titled "Sparks of Artificial General Intelligence," claiming it was a step toward achieving artificial general intelligence (A.G.I.) that can match human intelligence🤖.

• The research paper suggests that current AI systems, like OpenAI's GPT-4, can generate human-like answers and ideas NOT explicitly programmed into them.

• The researchers requested GPT-4 to write a mathematical proof that there were infinite prime numbers in a rhyming manner, to write a letter in the voice of Mahatma Gandhi, addressed to his wife, advocating for an electron as a U.S. presidential candidate and they asked the AI to generate a dialogue between Socrates and Aristotle that explored the misuses and dangers of L.L.M.s and the AI managed to correctly answer each of these questions ‘correctly’ 🤯

• Most impressively, they asked it to provide instructions on how to balance a book, eggs, laptop, bottle, and nail in a stable manner it provided a perfectly detailed answer, proving that it had a good understanding of the physical world. It suggested placing the book as a base, the 9 eggs in 3 rows of 3 each with space between them, the laptop with the screen facing down, followed by the bottle, and finally the nail on the top 👀.

• However, some AI experts criticize the Microsoft paper, considering it opportunistic and lacking rigorous scientific evaluation. The behavior of systems like GPT-4 can be inconsistent, and it is unclear if the generated text reflects human reasoning or common sense. The claims made in the paper could not be independently verified, as the tested version of GPT-4 was not the same as the public version💸.

Also, does that mean we can now have conversations with the dead?

Here’s an in-depth, extremely interesting article on this paper by Cade Metz from the New York Times.

A.I. Showcase: The ChatGPT App!

This week’s tool isn’t a particularly new one but it’s one that you must have. Finally, after several fakes available online and on the iOS stores, OpenAI has finally launched its official ChatGPT app! 📲

Yes… you heard it right! Just 2 hours before writing this, OpenAI released an iOS application (android releasing soon) for ChatGPT. They hadn’t previously hinted towards the application but given the popularity of ChatGPT, it was inevitable 📈. Here’s how to use it and a summary of its features!

Instructions:

• Download it here!

• Sync your chat history with the web by simply signing in

• *Optional - Activate the voice input (this is possible due to OpenAI's Whisper speech recognition model.)

• Talk to it like you always have texted it!

Information:

• It is available for both iPhones and iPads and can be downloaded from the App Store.

• OpenAI plans to initially roll out the app in the US and expand to other countries in the coming weeks📆.

• OpenAI previously introduced a premium subscription, ChatGPT Plus, which provides priority access and responses generated using the latest language model, GPT-4, at a cost of $20 per month💰.

• The app is expected to compete with Microsoft's Bing app, which offers access to its own GPT-4-powered chatbot, and may attract users away from it.

• The official ChatGPT app aims to deter users from fake or spam apps claiming to offer access to the chatbot on mobile🔏.

• Privacy and misinformation remain concerns, with OpenAI recently adding the option for private conversations and cautioning users against sharing sensitive information on the app.

JARVIS is getting closer and closer to reality…

Here’s the OpenAI blog, Twitter thread, and an article from The Verge for more details!

AI Art of the Week

This week, we have @chenheyuan27. With a unique style and overflowing creativity, @chenheyuan27’s work captivates the audience. We especially loved this ‘knock knock, who's outside the door?’ series, which is not only fun but shows off a great amount of detail.

Can you guess who’s outside all these doors?

You can find them on their Instagram: @chenheyuan27

See you soon!

-AI BitsI