The Download: AI Agent Infrastructure and OpenAI’s Expanding Ambitions

By Charlotte Jee — August 5, 2025

---

These protocols will help AI agents navigate our messy lives. A growing number of companies are launching AI agents that can perform tasks on your behalf—such as sending emails, creating documents, or editing databases. However, early reviews of these agents have been mixed because they struggle to interact seamlessly with the many components of our digital lives. Anthropic and Google are among the organizations working to solve this problem. Over the past year, both have introduced protocols aimed at defining how AI agents should communicate with each other and the external world. If successful, these protocols could form a crucial part of the infrastructure needed to make AI agents genuinely useful. Read more about these protocols. OpenAI operates under a dual mandate: it is both a tech giant focused on products—most notably ChatGPT, which reportedly receives 2.5 billion messages daily—and a research lab dedicated to creating artificial general intelligence (AGI) that benefits all humanity. My colleague Will Douglas Heaven recently interviewed the two key figures at OpenAI responsible for these research ambitions. Their insights reveal how OpenAI approaches product safety, what AGI truly means, and much more. Read the full story here. This story is from The Algorithm, MIT Technology Review’s weekly AI newsletter. Sign up here to receive it every Monday.

---

The must-reads

Here are today’s most important and fascinating technology stories:

• OpenAI is adding mental health guardrails to ChatGPT

OpenAI plans to make ChatGPT give less direct advice and encourage users to take breaks during lengthy conversations.

- NBC News coverage

- The Verge on AI mistakes in medicine

- OpenAI’s research on ChatGPT’s emotional impact

• The US wants to build a nuclear reactor on the moon

The US aims to beat Russia and China in deploying lunar nuclear power.

- Politico report

- NASA’s recent lunar mission failure

- Nokia’s first cellular network on the moon

• How to live forever (or at least get rich trying)

The booming longevity research sector is full of fascinating characters.

- New Yorker feature

- Longevity clinics selling unproven treatments

• Welcome to Silicon Valley’s ‘hard tech’ era

The focus is shifting from consumer software to large military contracts.

- NYT coverage

- MIT Technology Review on military AI

• There’s a big problem with the Gulf’s trillion-dollar AI dream

Building data centers in a water-scarce region raises sustainability concerns.

- Rest of World report

- US desert data center boom

- Google’s energy usage scaling

• Tesla’s board awarded about $30 billion of shares to Elon Musk

The board emphasized retaining Musk as critical.

- Financial Times report

- WSJ on CEO pay trends

• What happens if you respond to those scam job texts?

The exploitation can be surprisingly bizarre.

- Slate article

• Why there’s so much uproar over Vogue’s AI-generated ad

The controversy highlights ongoing debates about AI use in creative industries.

- TechCrunch coverage

• Earth’s core seems to be up and leaking out of Earth’s surface

This discovery challenges long-held geological assumptions.

- Quanta Magazine

- MIT Technology Review on volcanic glass brain

• Could lasers help us see inside people’s heads?

Promising new imaging methods face clinical hurdles.

- IEEE Spectrum

---

Quote of the day

"Hate it! Don't want anything to do with it."

>

— Weezy Simes, a 27-year-old florist, on AI

---

One more thing

What happened to the microfinance organization Kiva?

Since 2005, Kiva has connected lenders in wealthier communities with entrepreneurs worldwide, aiming to help people help themselves. But since August 2021, lenders have noticed essential information becoming harder to find, raising concerns that Kiva is now more focused on profit than impact. Read the full story. — Mara Kardas-Nelson

---

We can still have nice things

A place for comfort, fun, and distraction to brighten your day:
• I want this guy to draw my portrait.
• Highly recommend making these lemongrass chicken lettuce wraps. So tasty and easy!
• This encyclopedia teaches you about ancient gods and forgotten deities worldwide.
• Some of the architecture in Iran looks breathtakingly beautiful.

---

Source: Originally published at MIT Technology Review on August 5, 2025.

---

Frequently Asked Questions (FAQ)

AI Agent Infrastructure and Protocols

Q: What are AI agents and why are protocols needed for them? A: AI agents are systems designed to perform tasks on behalf of users. Protocols are essential for them to communicate effectively with each other and with the external digital world, ensuring seamless interaction and functionality across different systems. Q: Which companies are developing protocols for AI agents? A: Companies like Anthropic and Google are actively developing and introducing protocols to standardize AI agent communication. Q: What is the goal of these AI agent protocols? A: The goal is to create a crucial part of the infrastructure that will make AI agents genuinely useful by enabling them to navigate and interact with the complexities of our digital lives.

OpenAI's Ambitions and AGI

Q: What is OpenAI's dual mandate? A: OpenAI operates with a dual mandate: to be a leading technology company focused on products like ChatGPT, and a research lab dedicated to developing artificial general intelligence (AGI) that benefits all of humanity. Q: What does AGI mean in the context of OpenAI's research? A: AGI refers to artificial intelligence that possesses human-like cognitive abilities and can understand, learn, and apply knowledge across a wide range of tasks, aiming to benefit humanity. Q: What aspects of OpenAI's research are being highlighted? A: The article mentions insights into OpenAI's approach to product safety, their definition of AGI, and broader research ambitions.

Mental Health and AI Interactions

Q: How is OpenAI addressing mental health concerns with ChatGPT? A: OpenAI is implementing mental health guardrails, which include providing less direct advice and encouraging users to take breaks during extended conversations with ChatGPT.

---

Crypto Market AI's Take

The advancements in AI agent infrastructure, particularly the development of communication protocols, are foundational for the next wave of intelligent automation. As AI agents become more sophisticated, their ability to seamlessly interact with diverse digital systems will be critical. This is particularly relevant in the cryptocurrency space, where automation and efficiency are paramount. Our platform, Crypto Market AI, leverages AI to provide cutting-edge market analysis and trading tools. Understanding these infrastructure developments helps contextualize how AI will further shape financial markets and the broader digital economy, including areas like secure digital asset management.

---

More to Read:

• AI Agents: Capabilities, Risks, and Growing Role
• The Future of AI in Cryptocurrency Trading
• Understanding OpenAI's Approach to Artificial General Intelligence