SunBrief#38: OpenAI and Anthropic Criticize xAI’s Safety Practices Amid Controversies

Welcome to the SunBrief

Today in SunBrief 🌞

• Grow your technical skills with hands-on learning

• OpenAI Launches ChatGPT Agent

• Stock Updates

• Meta Settles $8B Lawsuit Over Privacy Failures

• AI Highlights of the Week

• OpenAI and Anthropic Criticize xAI’s Safety Practices Amid Controversies

• Too Important to Miss

Grow your technical skills with hands-on learning

Looking to stand out in a competitive job market?

Codecademy can help you build in-demand technical skills and prepare for top industry certification tests from AWS, Microsoft, CompTIA, and more.

📚 Smarter with AI readers can save 15% on a year of Codecademy Pro and get access to:

• ✅600+ interactive courses

• ✅Step-by-step career guidance

• ✅Industry certification prep

Try hands-on learning and get ready to unlock new career opportunities — just use code SKILLUP15 at checkout for 15% off. 🎯

OpenAI Launches ChatGPT Agent

The Article: OpenAI Launches ChatGPT Agent to Perform Tasks Like Managing Email and Files

OpenAI has unveiled its new "ChatGPT agent," a feature that allows the AI to autonomously use your web apps, including email and file management. This marks a significant expansion of ChatGPT’s capabilities.

Key Points:

• ChatGPT Agent: The AI can now browse the web, interact with apps, and perform tasks like sending emails, filling out forms, and creating files without direct user input.

• Pro Features: The agent is available to $200/month "Pro" subscribers, with lower-tier plans getting access soon. It can work on up to 400 tasks per month.

• Safety Measures: OpenAI has added strict safeguards, including user confirmations for sensitive tasks and active supervision, to ensure the AI does not perform high-risk actions.

Why It Matters:

The launch of ChatGPT agent changes the role of AI from a question-answering tool to a task-doing assistant. It enhances productivity by managing email, organizing data, and performing complex tasks, positioning OpenAI as a major player in AI-driven automation for professionals.

Stock Updates

The Smartest AI Library on the Internet🎉

Explore a world of 2789+ AI Resources from 97+ AI creators on LinkedIn, packed with actionable tutorials and diverse prompts!

Meta Settles $8B Lawsuit Over Privacy Failures

The Article: Zuckerberg Settles $8B Privacy Lawsuit Over Facebook Data Scandal

Meta and its executives, including Mark Zuckerberg, have settled an $8 billion lawsuit over privacy violations, stemming from the Cambridge Analytica scandal and failures to meet FTC privacy commitments.

Key Points:

• Privacy Breaches: Investors accused Meta leaders of ignoring privacy safeguards, leading to a $5B FTC fine in 2019.

• Settlement: The trial was adjourned after a quick settlement, though details weren’t disclosed. The lawsuit sought reimbursement from executives' personal wealth.

• Missed Accountability: Critics argue the settlement avoids full transparency on Meta’s privacy practices.

• Zuckerberg Avoids Testifying: The deal prevents Zuckerberg from facing questions under oath, sidestepping further scrutiny.

This settlement ends a significant legal battle but leaves privacy concerns unaddressed.

AI Highlights of the Week

• US passes first major national crypto legislation
  The US has passed its first major national crypto law, the Genius Act, focused on regulating stablecoins — cryptocurrencies backed by assets like the US dollar.

  The bill, backed by Donald Trump, requires stablecoins to be fully backed by low-risk assets. While seen as a win for the crypto industry, critics warn it could pose risks by legitimizing digital assets without enough consumer protections.
  

• Netflix Uses GenAI, Debuts First AI-Generated Scene

  Netflix has begun using generative AI in its original content, debuting its first AI-generated scene — a building collapse — in the Argentine show "El Eternauta."

  Co-CEO Ted Sarandos said the scene was made 10x faster and cheaper than traditional VFX, and emphasized AI as a tool to enhance creativity, not replace it.

  Netflix also uses GenAI for search, personalization, and plans to launch interactive ads later in 2025.
  

• Human Beats AI in World Coding Championship
  Polish programmer Przemysław Dębiak defeated a custom OpenAI model in a grueling 10-hour coding marathon at the 2025 AtCoder World Finals in Tokyo. Dębiak, a former OpenAI employee, narrowly outscored the AI in a head-to-head contest focused on complex heuristic optimization. The victory, hailed as a temporary win for humanity, highlights the growing prowess of AI in coding competitions.
  

• Google Adds AI-Powered Local Business Calling to Search
  Google Search now offers AI-powered calling to local businesses, allowing users to check prices, book appointments, or confirm availability without making a call themselves.

  The feature is available in 45 U.S. states and offers higher usage limits for AI Pro and AI Ultra subscribers. Calls are recorded for quality, and businesses can opt out.

  This move boosts Google’s agentic AI push as competition from AI browsers like Perplexity’s Comet increases.

OpenAI and Anthropic Criticize xAI’s Safety Practices Amid Controversies

The Article: OpenAI, Anthropic Call Out Elon’s xAI’s for Reckless AI Safety Practices

AI safety experts from OpenAI and Anthropic have raised alarms over xAI's "reckless" safety culture, following incidents involving its AI model, Grok.

Key Points:

• Safety Issues: Researchers criticize xAI for not publishing safety evaluations for Grok 4, which recently made headlines for offensive behavior, including antisemitic comments.

• Lack of Transparency: Experts call for more transparency in AI safety testing, stressing that unaddressed issues could have serious consequences.

• Regulatory Push: The criticisms may lead to calls for legislation requiring AI companies to release safety reports and follow stricter protocols.

• Musk’s Contradiction: Despite Musk's previous support for AI safety, xAI’s practices seem to conflict with industry standards, raising concerns about public trust.

Too Important to Miss

Copilot Can Now See Your Screen

Uber to Launch 20,000 Autonomous Taxis

Amazon launches AI agent-building platform

OpenAI partners with Google Cloud for ChatGPT

Perplexity Aims to Pre-Install AI Browser on Phones

Firefly Adds AI Sound Effects, Avatars, and Advanced Video Controls

Last Week’s Poll Result

• Do you think the competition for AI talent between companies like OpenAI, Meta, and xAI will lead to better advancements in AI?

  Yes, 64.71%. Maybe, 23.53%. No, 11.76%.

• Do you think OpenAI’s AI-powered web browser can successfully compete with Google Chrome?

  Yes, 57.14%. Maybe, 21.43%. No, 21.43%.

• How concerned are you about the potential adoption of AI models like Grok 4, given past controversies surrounding xAI?

  Yes, 66.67%. Somewhat, 33.33%. No, 00.00%.

Feedback

Know someone who may be interested?