Good morning, I'm your AI Brief anchor. Here's what's happening in AI today, Monday, June 29, 2026.
OpenAI Gives ChatGPT a Memory
Our top story today: OpenAI is rolling out a long-awaited memory feature for ChatGPT. This major update is designed to end the frustration of repetitive conversations, allowing the AI to remember key details and preferences across all of your chats.
Starting today, ChatGPT can learn things you tell it, like your job, your writing style, or that your daughter loves jellyfish. The goal is a more personalized and efficient experience. Instead of reminding the AI of your project's specific requirements in every new session, it will retain that context, making follow-up questions much faster and more intuitive.
OpenAI says users will have full control over ChatGPT’s memory. You can view what it has remembered, tell it to forget specific information, or turn the feature off entirely. While the initial rollout is for Plus subscribers, the company plans to make it available to all users soon. This move marks a significant step toward making AI assistants feel less like a tool you pick up and more like a partner that understands you.
EU Passes Landmark AI Regulation
Moving from product innovation to policy regulation, the European Parliament has officially passed the 'AI Trust & Transparency Act,' or ATTA. This is a landmark piece of legislation that will fundamentally reshape how artificial intelligence is developed and deployed across the European Union, with ripples expected worldwide.
The centerpiece of the new law is a requirement for mandatory third-party audits for any AI system deemed "high-risk." This includes AI used in critical sectors like healthcare, law enforcement, and hiring. Companies operating in the EU will now be legally required to prove their systems are safe, transparent, and unbiased before they can be released to the public.
Many are calling the ATTA the AI equivalent of GDPR, the EU’s sweeping data privacy law that set a global standard. Supporters are hailing it as a crucial step for building public trust and ensuring accountability. However, some industry groups have raised concerns about the compliance burden, arguing it could stifle innovation for smaller startups. Either way, the era of self-regulation in Europe is officially over.
SynthAI Breach Exposes 5 Million Users' Private Data
And that focus on trust is being tested today, as AI solutions provider SynthAI has confirmed a massive data breach. The incident exposed the sensitive information of over five million users, including private datasets and confidential prompts.
The company traced the breach back to a misconfigured cloud storage bucket, a simple but devastating security lapse. The exposed data includes proprietary business plans, personal creative projects, and research data that users had uploaded to train custom AI models. This isn't just about names and email addresses; for many users, this is the intellectual property at the core of their work.