AI This Week

The U.S. government has lifted export controls on Anthropic's most advanced AI models. The company confirmed it received notice that the Department of Commerce lifted the ban on Claude Fable 5 and Mythos 5, and pledged to begin restoring access the following day. The ban had abruptly barred foreign nationals — including Anthropic's own non-U.S. employees — from accessing its newest models, citing national security concerns. The trigger? Amazon found a jailbreak — a way around Fable's guardrails — though Anthropic argued the workarounds were "simple" and existed in other publicly available models. Anthropic ultimately trained a new safety filter that blocks the reported workaround more than 99% of the time. The reversal also comes as Chinese open-source models are beginning to show similar cyber capabilities to those of major U.S. AI labs. The White House had also requested OpenAI limit the release of its upcoming GPT 5.6 model to a small number of government-approved partners.

Music streaming platform Tidal is drawing a hard line against artificial intelligence in music. The service announced that artists uploading music wholly or substantially created using AI will be required to label it — and those songs won't be eligible for royalties. The policy goes into effect July 15 and also applies to Tidal's independent artist franchise, Tidal Upload. Fully AI-generated music will be identified and tagged with an "AI" badge, and those tracks will not be able to collect royalties or qualify for direct-to-fan sales. Tidal will also ban AI-generated music associated with "fraudulent activity," including songs impersonating established artists. The policy was sparked by listener complaints and a flood of fully AI-generated content that often impersonates existing artists for financial gain, according to Tidal EVP Tony Gervino. Tidal is the first major platform to test whether removing the financial incentive actually reduces AI music volume. The policy joins similar efforts from Spotify, Apple Music, Deezer, and Qobuz.

Austria is urging the EU to host Anthropic's AI systems within its borders following a U.S. export-control order restricting foreign access to its advanced models. The move signals a dramatic escalation in the transatlantic AI rivalry. On June 12, the U.S. issued an export-control directive blocking all foreign nationals from accessing Anthropic's Fable 5 and Mythos 5, prompting a global shutdown of those models. The U.S. Commerce Department has since allowed limited restoration, but the incident underscores geopolitical risks tied to centralized AI development. In a letter to EU Technology Commissioner Henna Virkkunen, Austria's State Secretary for Digitalization Alexander Proell wrote it was important that Europe was not cut off from major innovations. He did not say how the step could be taken and acknowledged there would be skepticism about whether it was possible. Earlier this month, the European Commission proposed laws to boost domestic cloud, AI, and semiconductor industries and cut reliance on U.S. Big Tech. Anthropic did not immediately reply to a request for comment.

The Trump administration has ordered a dramatic new gatekeeping process for OpenAI's latest AI model. The administration asked OpenAI to stagger the release of GPT-5.6 over security concerns, limiting early access to a small group of partners while the federal government approves customers one by one. CEO Sam Altman announced this during an internal Q&A session on Wednesday. The request comes after the administration placed an export control order on Anthropic, leading the company to pull its most advanced models, Mythos and Fable — which raised fears in Washington over their advanced cybersecurity capabilities. OpenAI and the administration view GPT-5.6 as "on par" with Mythos. The push for a phased release came out of talks with the Office of the National Cyber Director and the Office of Science and Technology Policy. Even after OpenAI shared its plans for the limited release with senior government officials, Altman received a call from Commerce Secretary Howard Lutnick, who warned the company against moving forward without sign-off from more agencies. As it stands, there is no transparent, consistent framework for regulating AI.

In a letter dated June 10 addressed to U.S. Senate Banking Committee Chair Tim Scott and Ranking Member Elizabeth Warren, Anthropic alleged that operators affiliated with Alibaba and its AI research division, Alibaba Qwen, executed a systematic campaign to illicitly harvest Claude's capabilities. Anthropic said the operators carried out 28.8 million exchanges with its models using roughly 25,000 fraudulent accounts between April 22 and June 5. Distillation is the practice of feeding carefully constructed queries to a frontier AI model, collecting its responses, and using those responses to train a cheaper rival system that approximates the original's capabilities. The campaign targeted what Anthropic called Claude's most valuable capabilities, including agentic reasoning, software engineering, and long-horizon tasks. The accusation marks the first time Anthropic has named a major Chinese technology conglomerate as the source of a distillation attack. The Alibaba campaign alone exceeded the combined volume of all three earlier efforts against DeepSeek, MiniMax, and Moonshot AI. Anthropic wrote that in proceeding with its distillation attacks, Alibaba "ignored the Trump Administration's warnings." Alibaba's American depositary receipts fell more than three percent on the news. Alibaba has not publicly responded.

Anthropic has launched an AI agent inside Salesforce's Slack app that can work alongside employees in group chats, deepening its push into the enterprise market. Called Claude Tag, the tool lets users summon the AI agent into a Slack thread by typing "@Claude," where it can read conversations, break down tasks, and proactively flag relevant updates across an organization without being asked. Claude Tag works with Opus 4.8 and replaces the older Claude in Slack app, with admins given 30 days to opt in before the prior Slack experience switches over on August 3, 2026. The agent retains context over time and is available in beta for Claude Enterprise and Claude Team customers, with plans to expand to other platforms. The move comes as strong early demand from businesses has helped push Anthropic's valuation to $965 billion, surpassing that of OpenAI. Administrators can tightly control the data and tools that Claude Tag can access in each channel.

Google and A24 have struck an AI research partnership, with Google investing roughly $75 million in the independent studio to work with its DeepMind unit on new AI-powered technologies for filmmakers — a sum in line with what Thrive Capital invested in A24's last funding round. The deal gives A24 access to DeepMind's research and infrastructure while researchers build out new workflows, but crucially does not give Google access to A24's content library or data. A24 partner Scott Belsky, who leads technology division A24 Labs, said the partnership differs from other AI deals because developers have mistakenly advertised AI as a way to make films cheaper and faster. His division is developing AI-generated storyboard applications, a reimagination of production that filmmakers like Martin Scorsese have already endorsed. The deal reflects Hollywood's turbulent AI era — Disney sued AI companies while briefly licensing its characters to OpenAI, Lionsgate expanded its Runway AI partnership, and Netflix acquired Ben Affleck's AI startup InterPositive.

Getty Images shares surged as much as 200% on June 22 after announcing a multi-year display partnership with OpenAI that puts licensed photos inside ChatGPT. The agreement allows Getty Images content to appear across ChatGPT search and discovery experiences, expanding the use of licensed visual content in AI-generated responses. Getty's arrangement is more guarded — it gives ChatGPT access to images for display and attribution, while keeping those assets out of OpenAI's model weights. The release disclosed no financial terms, no revenue split, and nothing about whether Getty's images may be used to train future OpenAI models. The deal is notable as it comes despite Getty Images losing its high-profile copyright infringement case against British AI developer Stability AI in November 2025. OpenAI has signed content arrangements with a long list of publishers, from News Corp to the Financial Times and Axel Springer. Getty is now the most prominent name to cross from litigation to licensing, and the first major visual-content owner to do so on this scale.

‍

AI image-generation pioneer Midjourney has stunned the tech world with a radical pivot. On Wednesday, Midjourney announced the establishment of Midjourney Medical, which it admitted was a bit out of left field. To promote the tech, it claims to be opening a spa in San Francisco where guests will step "into a shallow pool of golden light," before being lowered into a tank where ultrasound sensors bombard their bodies to produce AI-assembled, MRI-like images. Midjourney Medical is building a 60-second full-body ultrasonic scanner it claims rivals MRI at nearly 100 times the speed, with the first scanner-equipped spa set to open in San Francisco in 2027. Rather than selling people on going to a medical facility, Midjourney is wrapping the scanner inside a spa with hot tubs, saunas, cold plunges, and pools of golden light, open 24/7. "We think it's completely possible that with enough early imaging in the future, the world could avoid 30% of all deaths and 50% of all healthcare costs," the company said. The bootstrapped AI image lab is building these MRI-grade scanners with no outside investors and no full FDA sign-off yet. The stated goal is a fleet of over 50,000 scanners worldwide by 2031, capable of a billion scans a month.

ChatGPT commanded over 50% of the AI assistant market until January 2026, but by the end of May that figure had fallen to 46.4%, according to Sensor Tower's State of AI Report. The throne is wobbling. Gemini now holds 27.7% of the market, while Claude accounts for 10.3%. Grok, Perplexity, DeepSeek, and Meta AI each hold less than 5%. ChatGPT became the fastest app ever to reach 1 billion monthly users. But dominance is fading fast. OpenAI's deal with the U.S. Department of Defense in February triggered a measurable spike in uninstalls — suggesting brand trust and values alignment matter to users, not just features. Anthropic's Claude has gained a strong reputation for productivity use cases and is closing in on ChatGPT's user-retention rate. Thirteen percent of Anthropic's users are paying for a subscription plan, a conversion rate that leads the field. In H1 2026, people are on pace to spend over $4.2 billion on AI apps, compared to $1.83 billion in H1 2025 — a jump suggesting the industry is shifting from pure growth toward monetization.