Meta Scoops Up Apple Researchers As OpenAI, Citi & Ant International Boost AI

During moments of curiosity-fueled daydreaming, we all dream of having a robotic sidekick handle the mundane chores, right? We imagine having someone book the flight tickets, prepare business reports, or even plan your weekly breakfast schedule—without endless Google tabs and to-do lists.

Only to remember—it’s not real.

Well, this week, that long-standing sci-fi dream edged closer to reality. As always, it’s been a high-voltage week in the world of Artificial Intelligence (AI). We saw AI leader OpenAI launching its much-awaited ChatGPT agent and Citi group teaming up with Ant International to launch an AI tool—and, of course, Meta’s bounty hunt for AI talent.

Let’s break it all down!
 

OpenAI Launches ChatGPT Agent For Complex Real-World Tasks

OpenAI has officially rolled out its agentic upgrade to ChatGPT, allowing the chatbot to not only think but also act, using a virtual computer to complete complex tasks from start to finish.

The ChatGPT agent can now handle requests like analyzing competitors and creating slide decks, booking events, summarizing inboxes, and even planning meals by autonomously browsing the web, running code, and accessing connected apps like Gmail or GitHub.

At the heart of this upgrade is a unified system that merges its earlier tools—Operator and Deep Research—into a single model, capable of shifting between text-based and visual web interaction, API calls, and terminal commands.

Users can activate the feature through ‘agent mode’ under the tools dropdown. The rollout starts with Pro, Plus, and Team users, with Enterprise and Education users gaining access in the coming weeks.

ChatGPT Agent has demonstrated state-of-the-art performance on a range of benchmarks, including DSBench (data science), SpreadsheetBench (real-world spreadsheet tasks), and BrowseComp (web tasks), outperforming existing models and, in many cases, human experts.


According to OpenAI, “This launch is just the beginning,” as future updates aim to make the agent more polished, flexible, and widely available.

Meanwhile, the organization’s nonprofit side has been under the lens—with some important feedback now shaping its future direction.
 

OpenAI Board Responds To Nonprofit Commission Report

Overshadowed by the news about the launch of ChatGPT Agent, the OpenAI Board of Directors issued a statement responding to the Independent Nonprofit Commission’s findings while evaluating OpenAI’s nonprofit arm. The Committee reviewed if the AI leader could better address systemic challenges and ensure the development of AGI benefits all humanity.

“We remain committed to listening, learning, and building in partnership with those on the front lines of change,” the Board stated, emphasizing a renewed focus to become a well-resourced, mission-aligned nonprofit through community-led initiatives.

Based on inputs from community leaders, philanthropic experts, technologists, and other stakeholders, OpenAI is expected to reshape its long-term vision, with the report Calling OpenAI “as uniquely positioned to be a force of good in this Age of Intelligence.”

While this move aligns with the demand for accountability, the financial world is also witnessing greater transparency thanks to the AI shift—especially when it comes to smarter risk management.
 

Citi And Ant International Build Debut AI Tool To Cut FX Hedging Costs

In a joint move bridging fintech and traditional banking, Citigroup and Ant International announced a pilot AI-powered tool to help businesses optimize foreign exchange (FX) risk management.

The initiative combines Citi’s Fixed FX Rates solution with Ant’s Falcon Time-Series Transformer, a proprietary forecasting model, to enable smarter hedging strategies.

Targeting e-commerce and aviation sectors first, the tool has already delivered tangible results: in the case of one major Asian airline, the tool reportedly slashed FX hedging costs by 30% during the pilot phase.

“The cost efficiency that can be achieved with AI-enabled FX hedging is real,” said Kelvin Li, GM of Platform Tech at Ant International.

The collaboration underscores a growing trend of AI integration in financial services, with banks like Morgan Stanley and Bank of America also deploying tailored AI assistants for both advisory and consumer-facing operations.

Ant International, a spin-off from Jack Ma’s Ant Group, is expanding its offering across industries, sectors, and regions, aiming to democratize smarter financial risk management tools for enterprises across the globe.

With the demand to make AI smarter, there’s one shortage that the industry is facing—AI talent. To address this gap, Meta is taking some serious steps.
 

Meta Hires Ex-Apple AI Talent For Superintelligence Labs

Meta has quietly recruited two prominent AI researchers from Apple—Mark Lee and Tom Gunter—for its Superintelligence Labs team. The duo previously worked under Ruoming Pang, Apple’s former head of Foundation Models, who also jumped ship to Meta earlier this month in a multi-million-dollar move.

CEO Mark Zuckerberg is personally leading the charge to build a world-class AI division and has pledged “hundreds of billions of dollars” to expand Meta’s compute and R&D infrastructure.

These hires are part of Meta’s aggressive expansion of its Superintelligence initiative, which aims to develop AI systems capable of surpassing human intelligence.

While Apple has not commented on the departures, these exits highlight the growing competition in the AI talent war. As technology giants race to build foundational models and next-gen agents, access to top-tier talent is getting scarce and competitive.

Do you think OpenAI’s new ChatGPT agent will be a big leap forward for AI assistants? Will Meta’s onboarding of Apple AI researchers help it take the AI spotlight?

Let us know in the comments below!