We Must Regulate AI, FTC Chair

PLUS: ChatGPT Chops Chegg, Open Source Is The Answer

Good morning! In today's news, we’re covering FTC Chair, Lina Khan, urging AI regulation to prevent unwanted risks, open-source AI models key role in allaying bias and deception concerns, and ChatGPT's popularity causing Chegg's stock to plummet as students favor AI-based homework assistance.

Let’s get to it!

Top News & Why It Matters

| “We Must Regulate AI,” FTC Chair Khan Says

FTC Chair Lina Khan recently shared her thoughts on AI regulation in a New York Times op-ed titled "We Must Regulate A.I. Here's How." Khan emphasizes the need to protect against possible AI risks such as market dominance by big tech companies, collusion, and increased fraud and privacy violations.

She refers to the "Web 2.0" era as a cautionary tale to be mindful of when managing AI's expansion, citing invasive surveillance and loss of privacy that resulted from the growth of tech firms. Khan's goal is to ensure that history doesn't repeat itself with AI, while still maintaining an environment that promotes innovation. arstechnica

Why does this matter?

It's essential to strike a balance between regulation and innovation in the AI industry. Too little oversight could lead to negative consequences, but excessive restrictions might stifle progress and competitiveness.

| Why Open Source Is Essential To Allaying AI Fears, According To Stability.ai Founder

Stability.ai founder Emad Mostaque insists that open-source AI models are essential to addressing fears about biased or deceptive AI systems. Not knowing how powerful programs like GPT-4 and DreamStudio work could be a significant barrier to trusting them with sensitive data, according to Mostaque. He believes that making AI systems open-source can help tackle issues such as bias, unethical output, and copyright infringement. zdnet

Why does this matter?

An increased demand for open-source AI could lead to a greater focus on collaboration among researchers and developers, potentially accelerating the development and adoption of AI technologies in various industries. However, the balance between transparency and intellectual property protection will remain a challenge that must be navigated carefully.

| ChatGPT's Popularity With Students Slices Chegg's Stock Nearly In Half

ChatGPT's surge in popularity has caused a dramatic drop in Chegg's stock value as students increasingly turn to AI for homework assistance. Chegg, a subscription-based service for academic help, has seen a decline in new subscribers, with CEO Dan Rosensweig stating that ChatGPT is affecting Chegg's growth rate. Chegg's stock plunged nearly 50% ahead of the market open on a recent Tuesday.

In response to the growing demand for AI-based tools, Chegg is developing an AI project called CheggMate, which will incorporate GPT-4 technology with Chegg's content to create personalized learning experiences for users on its platform. zdnet

Why does this matter?

Chegg's plummeting stock value signals the rapid shift in demand from traditional educational resources to AI-powered tools like ChatGPT. As AI applications continue to evolve and offer more advanced capabilities, Chegg and similar companies will need to adapt and innovate, or risk becoming obsolete in the ever-evolving edtech landscape.

Mastering Artificial Intelligence

| Our Top Picks

✅ ChatGPT for Search Engines: ChatGPT extension with code syntax highlights and search engine results. (website)

✅ EzMail.AI: AI email drafter drafts personalized responses 10X faster. (website)

✅ Jusi: AI tool automates technical specs and talent search, saving time/cost. (website)

✅ ThankYouNote.app: A service where AI is used to help write the best thank you notes for any occasion. (website)

✅ AYLIEN News API: AI news intelligence platform for real-time global news analysis. (website)

| Lunch & Learn

📝 Text: FAQs about Superintelligence, by Scott Alexander (link)

🎧 Audio: “Practical AI” by Changelog Media (link)

📽️ Video: Intro to Artificial Intelligence (link)

What Else Is Happening

| More News

🗞️ AI can’t replace human writers. (source)

🗞️ Four investors explain why AI ethics can’t be an afterthought. (source)

🗞️ GPT-4 cheat sheet: What is GPT-4 & what is it capable of? (source)

🗞️ Your phone contains your most personal information. Here's how to keep it safe. (source)

🗞️ How generative AI can help boost your productivity. (source)

🗞️ Google used AI to make a puzzle game that rivals Wordle. (source)

🗞️ This AI-powered exoskeleton does the heavy lifting so you don’t have to. (source)

🗞️ Video: Geoffrey Hinton talks about the “existential threat” of AI. (source)

🗞️ ChatGPT scams are the new crypto scams, Meta warns. (source)

🗞️ Tech job opportunities remain strong despite layoffs — here’s why. (source)

🗞️ AI is just someone else’s computer. (source)

🗞️ AI-powered supply chain startup Pando lands $30M investment. (source)

| Best of AI Social Media

💬 AI Art: Star Wars Characters Imagined as Teapots (Twitter)

💬 AI graphics revolution driven by NVIDIA research. (Twitter)

💬 "Top 10 AI Productivity Tools: ChatGPT Amongst 1000 New Releases" (Twitter)

💬 AI Vision Assistant Learns Keto Diet, Identifies Foods, and Finds Recipe (Twitter)

💬 ChatGPT gives the same quality of answers as a certified senior psychologist (Reddit)

💬 If we don’t ‘pause’ A.I., CHINA WILL WIN. Here’s why. (YouTube)

| Research Review

Automated Scientific Discovery: From Equation Discovery To Autonomous Discovery Systems (Full Paper)

Stefan Kramer et al.

∑ Summary: The paper surveys automated scientific discovery approaches, addressing open issues and recent topics, while discussing autonomy levels in AI systems for achieving Nobel-quality scientific discoveries by 2050.

!? Practical implications: Non-obvious second-order effects of this research on industry or society could include exponentially accelerated scientific innovation, rapid advancements in a wide range of fields, and potentially reduced demand for human researchers in certain areas, as AI systems leverage their autonomy to make Nobel-quality discoveries.

Continual Reasoning: Non-Monotonic Reasoning In Neurosymbolic Ai Using Continual Learning (Full Paper)

Sofoklis Kyriakopoulos, Artur S. d'Avila Garcez

∑ Summary: This paper explores Continual Reasoning, a neural-symbolic system combined with continual learning methods, to improve accuracy in non-monotonic reasoning tasks.

!? Practical implications: This research could lead to advancements in AI applications that require non-monotonic and commonsense reasoning, potentially improving industries such as healthcare, finance, and manufacturing by allowing AI systems to make more effective decisions based on incomplete or changing information.

Written by Isaac R. Ward, Casey Clifton, and Alex Brogan.

Send us your feedback at [email protected]. We’re ready to listen.

Join the conversation

or to participate.