AI Tech Helper Weekly: Malta Bets Big—Free ChatGPT Plus for Every Citizen

It’s been a week of AI tools not just expanding, but becoming inseparable from daily life—for better or worse. From government partnerships pushing AI directly into the hands of citizens, to smarter agents now available on our phones, the pattern is this: AI isn’t just a tool—it’s fast becoming infrastructure. Let’s dig into what that means for how you’ll live, work, and manage your money.

TOP STORIES

Malta to Offer Free ChatGPT Plus Access to All Citizens

• Malta partners with OpenAI to give every citizen free access to ChatGPT Plus, the paid version of ChatGPT.
• The initiative is paired with an AI literacy course to teach practical skills and boost digital inclusion; rollout starts this year.
• Malta is the first country to offer this kind of nationwide AI access, signaling how governments might treat AI as a public good.
• Source

GPT-5.5 Arrives for Enterprises, Boosting Office Workflows

• Databricks is bringing OpenAI’s new GPT‑5.5 to its enterprise customers for corporate AI agents.
• This new model achieved 50% accuracy on Databricks’ “OfficeQA Pro” benchmark for complex office tasks—a jump over GPT‑5.4, with a 46% reduction in errors.
• GPT‑5.5 is now available to Databricks customers to power more accurate, reliable agent workflows.
• Source

ChatGPT Pro Users Get Personal Finance Tools—With Bank Data Integration

• OpenAI is testing a new personal finance feature for ChatGPT Pro users in the US, letting them securely connect bank accounts and get spending breakdowns.
• Users can ask ChatGPT questions based on their actual financial transactions—aiming to help people understand and control their money.
• Feature is rolling out as a limited preview to gather feedback on privacy and real-world use before broader release.
• Source

Shopee’s Sea Group: Agents Will Change How Software Gets Built

• David Chen, co-founder of Sea (parent of Shopee), shared his vision for “agentic” software development—apps that don’t just automate tasks, but build and improve themselves.
• Sea’s engineering teams now use AI agents for real, production-scale products, not just experiments.
• This signals a shift toward AI “co-workers” shaping the future of tech work.
• Source

Codex Agents Now Run on Your Phone—Work From Anywhere

• Codex, OpenAI’s coding agent, is now available in preview inside the ChatGPT mobile app, letting users monitor and manage coding tasks remotely.
• Users can collaborate with Codex as it works across their computers—even while away from their desks.
• This bridges the gap for remote and hybrid work, and hints at agents running persistently (and independently) on your behalf.
• Source

ChatGPT Gets Smarter at Handling Sensitive Conversations

• OpenAI rolled out new safety features so ChatGPT can better recognize—and respond appropriately to—sensitive or risky conversations as they unfold.
• The update relies on new context-aware systems to catch subtle cues over longer exchanges, aimed at safeguarding users in vulnerable moments.
• These changes are now live and apply to all users interacting with ChatGPT.
• Source

THIS WEEK IN AI

If you had any illusions left that AI is a sideshow or “nice to have,” this week should put them to rest. What we saw was a shift: AI isn’t just showing up on more platforms—it’s becoming part of public infrastructure, deeply entwined with private life, and (yes) our wallets.

The Malta–OpenAI deal is one of those moments that will show up in future textbooks. A whole country treating advanced AI not as a luxury, but as a base service—like water, electricity, or high-speed internet. For Maltese citizens, the benefits are obvious: access to the best tools and the know-how to use them, with the government footing the bill. But the more provocative question is, what happens when a country’s “default” search box and productivity assistant is built and hosted by a US company? Is this a digital welfare leap, or a big experiment in technology dependency?

Meanwhile, Databricks bringing GPT-5.5 to the corporate front lines means office work is about to become much less about copy-paste and far more about managing (and trusting) autonomous software agents. The stats are compelling—fewer errors, greater accuracy—but if you’ve ever wrestled with “automated” processes at work, you’ll know: technology is only as good as its context, and trust is still earned, not given. Will companies really re-train their workflows, or will they keep their humans in the loop as brake pads for inevitable AI hiccups?

On the personal front, OpenAI’s new finance features make your bank statements as “chatty” as your calendar or inbox. It’s tempting to think of this just as another dashboard. But letting an AI read and analyze your financial life isn’t trivial—are you willing to “invite” a digital assistant into your money conversations? What safeguards, what value, what trade-offs? We need to interrogate these before the feature rolls out wider.

Also telling: agents and coding tools are now available on your phone, not just your office PC—meaning the boundaries between work time and personal time, between “using” and “being used by” software, will blur even further. Add in the ongoing improvements for sensitive conversations, and we see an ecosystem preparing for deep integration—in your company, your communications, your daily decisions.

The call to action: Try out what’s new, but don’t sleepwalk into reliance—especially with personal data. If you’re in Malta, are you excited or wary to have an AI at your fingertips, sponsored by your government? If you’re offered the new ChatGPT finance tool, will you connect your accounts? Write in, or challenge yourself: What’s your red line for AI convenience versus privacy and control?

MORE TOP STORIES

Codex on Windows Gets a Safer ‘Sandbox’ for Coding Agents

• OpenAI is adding a “sandbox” environment so Codex agents can safely run on Windows without dangerous system access or endless approval pop-ups for the user.
• This lowers risk when letting coding AIs work on files, and brings Windows up to par with Mac/Linux support.
• Expected to improve reliability for developers and companies using agent-driven coding workflows.
• Source

OpenAI Responds to npm Supply Chain Attack Affecting TanStack

• OpenAI detected a security issue with the TanStack npm package—part of a wider “Mini Shai-Hulud” supply chain attack—but found no user data was accessed or software altered.
• The company took steps to shore up defenses and communicated steps to prevent future incidents.
• If you use OpenAI’s tools, no action is needed; the breach did not reach production systems.
• Source

Lessons from OpenAI’s ‘Parameter Golf’—1,000+ Compete in AI Challenge

• OpenAI shared insights from its ‘Parameter Golf’ contest, which drew over 1,000 participants and 2,000+ submissions to tackle a tough machine learning problem.
• The challenge rewarded creative solutions with tight technical constraints, serving as a testbed for agent-assisted research.
• Expect future contests and more community opportunities as OpenAI crowdsources new AI approaches.
• Source

NVIDIA Engineers Use Codex AI Agents for Major Productivity Gains

• Over 40,000 NVIDIA employees now have access to Codex-powered agents, with engineering cycles sped up nearly 10x in some cases.
• Codex helps automate repetitive coding tasks and run complex experiments across internal teams.
• Large organizations are rapidly making AI co-pilots a standard part of the workflow.
• Source

AutoScout24 Scales Up Engineering with AI Workflows

• AutoScout24, Europe’s largest auto marketplace, rolled out Codex and ChatGPT to 2,000 employees to accelerate development, with some cycles cut from weeks to days.
• Over 1,000 “builder roles” are now using AI to improve code quality and expand automation.
• The story signals mainstream adoption for AI in large, non-tech companies.
• Source

ChatGPT User Growth Accelerates Across Ages, Genders, and Countries

• Q1 2026 data show ChatGPT consumer adoption growing rapidly, broadening across more age groups, more women (based on name inference), and more geographies.
• The strongest growth is in users with typically feminine names, and adoption deepened in new regions.
• OpenAI credits new tools, more languages, and falling access barriers for this spread.
• Source

ALSO THIS WEEK

• OpenAI launches global campus program for student leaders — Student clubs can join a new OpenAI campus network to promote AI literacy and projects worldwide. (Source)
• “Normies” try out AI-powered coding — WIRED explores whether everyday users can actually code with AI assistants. (Source)
• New AWS tools help train giant foundation models — Hugging Face explains building blocks for AI model training and serving on Amazon’s infrastructure. (Source)
• Altman faces cross-examination in OpenAI trial — Sam Altman is grilled on credibility as legal battle over OpenAI’s control continues. (Source)
• New architectures for “graph-enhanced” AI search — VentureBeat covers how companies move beyond simple vector search for better AI answers. (Source)
• Intercom (now ‘Fin’) launches an AI agent to manage other agents — A customer-service AI whose full-time job is supervising and correcting other AIs. (Source)
• The risk: AI is replacing experts it’s supposed to learn from — Companies worry AI adoption erodes the very expertise needed to train and check it. (Source)
• Raindrop releases tool for debugging AI agents — Developers can now locally debug agents for transparency and trust, using a new open-source project. (Source)
• Musk v. Altman trial enters final phase — High-profile trial prompts questions on OpenAI’s leadership and ambitions. (Source)
• Multi-agent system speeds up AI, slashes costs — RecursiveMAS claims to cut latency and token use in complex AI applications by 2.4x and 75%, respectively. (Source)
• Greg Brockman takes over all OpenAI products — Brockman consolidates control over OpenAI’s product lines in latest internal restructuring. (Source)
• Anthropic’s Claude aims to compete at the infrastructure level — New report highlights battle for control of “AI agent” orchestration, not just better models. (Source)
• Disco Elysium’s spiritual successor reviewed — A new game explores themes of forgiveness and the human condition amid post-AI society. (Source)
• Cerebras stock doubles after IPO; signals AI chip boom — AI chipmaker goes public, sending shockwaves through the supply chain and infrastructure sector. (Source)
• Mira Murati, ex-OpenAI CTO: Keep humans in the loop — Interview calls for keeping human insight central to AI progress—even as superintelligence gets closer. (Source)
• Musk v. Altman trial: closing arguments — Final legal arguments in the case that could alter OpenAI’s direction. (Source)
• Data readiness is key for AI in finance — MIT Technology Review explains why financial firms need robust context data for safe, useful AI. (Source)
• Why AI/data sovereignty matters in the age of autonomous systems — Sovereign control over data and models is a rising focus for enterprise AI buyers. (Source)
• AI chatbots leak real phone numbers — Reports emerge of Google AI surfacing people’s private contact info, with no easy fix yet. (Source)
• 60+ Trump allies want AI “vetted” before public use — Political letter urges pre-release testing and approval for the most powerful AI models. (Source)
• AI backlash now a real business risk — New polling shows public skepticism is growing against AI, and companies can’t ignore it. (Source)
• Sober, less social teens — Gen Z is drinking less, but also interacting differently, with possible links to digital-native lifestyles. (Source)
• Trump advisers fear China–Taiwan flashpoint — Behind closed doors, some urge focus on security risks from advances in tech. (Source)
• GOP: Crypto’s biggest donors sit out the midterms — Republican leaders frustrated by lack of crypto funding ahead of key races. (Source)
• Sen. Bill Cassidy loses Louisiana Republican Senate primary — Upset comes in a turbulent primary season. (Source)
• Hugging Face found hosting malware posing as OpenAI — 244,000+ infected downloads before removal raise new security concerns for AI developer tools. (Source)
• House Ethics panel investigates Rep. Chuck Edwards — Congressional investigation into workplace misconduct allegations. (Source)
• ChatGPT fuels A-grade boom at colleges — Universities see uptick in top grades, possibly tied to widespread use of AI assistance. (Source)
• Corporate perks vanish as economy shifts — Companies pull back on extra benefits amid cost pressures. (Source)
• Fed Chair Powell: a legacy of crisis navigation — Retrospective on the US central bank’s leadership through tumultuous times. (Source)
• Deloitte urges shift from basic AI features to “autonomous intelligence” — Firms told to go beyond chatbots and embrace agent-based automation for real value. (Source)
• US–China tech race: three new conflicts, new risks — Updates on the global race for AI dominance. (Source)
• Stock vs. bond market confusion as AI shapes the economy — Dueling market signals on growth, warnings, and inflation risk. (Source)
• Trump–Xi summit closes with uncertainty — Diplomatic relations remain in flux after headline meetings. (Source)
• Democrats target hidden Latino battlegrounds — Political strategists zero in on under-the-radar districts. (Source)
• The reality check on “imperfect data” for business AI — JBS Dev president busts myths about what’s required for real-world AI deployment. (Source)
• Laserfiche debuts AI assistants for business workflows — Cloud platform adds AI agents for more flexible automation, following compliance rules. (Source)
• AI shakes up HR compliance (but not everywhere) — Automation helps most HR functions, but critical areas still need human oversight. (Source)