Guard Fails 🛟🤖

What’s Appening: AI Protections, Claude Hacking, Digg, Nvidia Revenues

ANOTHER GOOD TRICK FROM NVIDIA

👀 TLDR

In this week’s What’s Appening, we cover the rush from AI companies to deliver parental controls to their chatbots. It comes as OpenAI is facing a lawsuit from the family of a teenager who took his life after allegedly being encouraged to do so by ChatGPT. We also have news on the rules around Alexis Ohanian’s new social media site, the AI glitches at Taco Bell, how just two customers are propping up the world’s most valuable company, as well as all the latest from 3Advance and Shift. ;)

🎬 Our Top Take

⛑️ 🤖 AI - OpenAI and Meta Scramble for Controls After Teen’s Suicide

Back in April, California teen Adam Raine decided to end his life. A terrible nightmare for any family, and one that is way too common in today’s world. In this case, Adam’s parents have put the blame firmly on ChatGPT. The Raine Family is suing OpenAI for wrongful death – the first recorded case of its kind against an AI company – alleging that the chatbot offered months of encouragement (on suicide) and an accelerator for self-harm. This was a ticking time bomb, and unfortunately, more stories like this will emerge as more and more people turn to ChatGPT for medical and psychological support. Like the introduction of social media, though, it’s innovation first, safety later. Especially in the United States, the guardrails are largely determined by the companies, not government policy. That said, OpenAI announced that the first parental controls are coming to ChatGPT “this month.” Apparently, parents will receive a message from ChatGPT if their teenager is in “acute distress”. Meta, whose AI bots have caused a host of other problems with minors (and adults), also said yesterday it is banning its chatbots from discussing suicide and self-harm with kids. Even free-speech absolutist Elon Musk says a kid-friendly “Baby Grok” is coming soon. It’s like everyone has kicked into action at once. Or the PR machines are simply responding to news of a major lawsuit. #GuardRails Read more here.

🗣️ Poll: Should AI bots have more guardrails for kids?

Login or Subscribe to participate in polls.

📲 App of the Week: StoryShots

With summer on the wane and the school year beginning, there are plenty of apps that can help students of all ages. For older kids, StoryShots is one of the best. It summarizes key points in books (textbooks, novels, etc.), provides free audiobooks, and can even help with more advanced learning by providing cliff notes. StoryShots has been around for a few years, but it has evolved over time, arguably becoming the best book summarizer app around. It’s free with ads, whereas the paid subscription – access to 300K books, more features, etc. – is a budget-manageable $29.99 per year. Get it for iOS here and Android here.

🧐 Stat of the Week: 39%

📈📉 Nvidia – Over One Third of Revenue from Just Two Customers

Nvidia released its quarterly earnings last week. The call tends to act as a bellwether for the health of the tech sector, AI industry, and stock market at large, so, thankfully, expectations were beaten again by the world’s most valuable company. Yet, the filings also showed that 39% of Nvidia’s sales came from two mystery customers. It’s not clear who Customer A and Customer B are – Customer A accounts for almost a quarter of sales alone – but experts have warned that having such a high concentration of sales in so few hands is risky. #MysteryCustomers Read more here.

🌮 Extra Sauce? Taco Bell is reportedly reconsidering its rollout of AI drive-thru tech after a series of blunders, some of which went viral. One video showed a customer crashing the system by ordering 18,000 water cups. More here.

🚑 AI SOS. 911 call centers are apparently overwhelmed due to low staffing numbers, so they are turning to AI for call handling help. As per the Taco Bell story above, what could possibly go wrong? Details here.

😎 Just Be Cool. Reddit founder Alexis Ohanian’s new social media platform, Digg, has one overarching rule – don’t be an asshole (his words). Digg was actually founded in 2004, but it’s being relaunched. Story here.

🛠️ Hacking Tools. Anthropic has admitted that hackers have used Claude to make large-scale cyber-attacks. Claude was even suggesting ransom amounts for the victims. Find out more here.

📍 Meanwhile at 3Advance

Our Digital Product Studio has kicked into top gear this year. As you might know, Paul is a Co-Founder of Shift, and the 3Advance team has taken on the mission to connect the global Irish. So connect we shall! ☘️ We’re excited to share a sneak peek of the app that will be launching next month here in Washington, DC. Paul recently made this announcement on LinkedIn, and a callout for interested Irish investors to give him a shout.

What’s Appening is brought to you by 3Advance, your AI-driven product development team that speaks your language, too. If you’re not into it, simply opt-out below. If you think a friend might be interested, please forward it along 🙏.

You made it to the end of this email 🥳

How did you like "What's Appening" this week?
1  |  2  |  3  |  4  |  5  |  6  |  7  |  8  |  9  |  10

Login or Subscribe to participate in polls.