Time for a catch-up!

Well, it has been a quieter week compared to the recent past, but it has not been without controversy with some key events around data protection and copyright issues, although in my view they have not been publicised very much.

What was in the news were further revelations about Meta effectively trying to corner the market through its aggressive campaign of buying up companies, poaching staff, or both. More details emerged about them offering key people within OpenAI a $100 million signing up bonus!! 😳

Yep, you read that correctly. $100 million in your bank before you’ve even done a day’s work.

Meta clearly thinks that it will make lots of money, as they also announced yet another planned infrastructure development valued at $29 billion, this time with their plan to fund AI campuses in the US. Given the fact that $26 billion of that is planned as debt, they need lots of income, just to cover the interest payments

However, like I said, although it’s been a relatively quiet week, there still a lot happening. So, let’s dive right in …

Big Sharks vs. small fish

Sticking on the theme of income and market value, Nvidia reached a record-high market cap of $3.76 trillion, overtaking Microsoft’s valuation, underlining the optimism in the market around Generative AI.

Again, to underline the market’s interest in AI infrastructure, the European Commission received 76 applications for its €20 billion AI gigafactory fund, aiming to bolster AI infrastructure across Europe.

Whilst there were not any really significant releases of new software or LLM models, OpenAI did announce that ChatGPT 5 would be released this summer. However, given the fact that we’re in July already you wonder why they can’t be more specific on the actual date? 🤔 However, the suggested features are impressive, with multimodal integration (text, image, voice), trillion-token context windows, and improved reasoning capabilities.

Money was still being spent though, with Grammarly spending $1 billion to buy Superhuman, thus combining their 40 daily users with a platform that helps users process 72% more emails per hour than traditional platforms. So, suggestions on the death of emails may be wishful thinking.

Legislation, policy and other news

So, this was the area this week which dominated the news for me, so much so that I had to think about where to start. Probably the best place is the EU AI Act and the concerns that not everything is as ready as it should be (anyone surprised!?).

As the EU AI Act prepares to take effect on August 2, major tech firms are requesting a 2‑year postponement due to missing compliance guidance, with a code of practice now not expected until the end of 2025.

Add to this the fact that Computer & Communications Industry Association (CCIA) Europe, backed by Sweden’s PM, is advocating for a delay in implementing the EU’s new AI Act to clarify compliance demands, a position echoed by leaders from Airbus, BNP Paribas, and others. They have urged Brussels to pause enforcement, citing complicated rules that could impede innovation, especially for smaller European firms.

On the other side of the Atlantic the U.S. Senate voted 99 - 1 to remove a proposed 10‑year freeze on state-level AI regulation from Trump’s “Big Beautiful Bill”, thus signalling strong bipartisan support for preserving state authority over regulation of AI.

However, in other news, following Anthropic’s win on the issue of Intellectual Property Rights connected to books it used to train it’s LLMs, Meta also won a similar lawsuit, with the Court again ruling that use of copyrighted works to train Llama falls under fair usage.

Jumping back across the Atlantic, the Danish Parliament passed a law that grants individuals ownership rights over their likeness and voice, empowering them to demand removal of unauthorised AI-generated content, which is arguably an effective measure against deepfake misuse. The law also includes protections for artists and allows parody/satire exceptions, with compliance fines scheduled for implementation this autumn.

Linked to the issue of Intellectual Property Rights is the access to relevant materials online. Whilst there was a lot of focus this week on Cloudfare’s new tool that blocks web-crawling bots, the key thing from the announcement was Cloudfare’s analysis of web traffic. For context, since Cloudfare handles roughly 20% of all internet traffic, when their CEO talks about AI reshaping the web, it’s worth listening. What he revealed was the change in the way the searches via AI are impacting the web generally.

10 years ago, Google crawled 2 pages of the web per visitor. Fast forward to 6 months ago and the statistics Google 6:1, OpenAI 250:1, and Anthropic 6,000:1. Now those figures have jumped even more to Google 18:1, OpenAI 1,500:1, and Anthropic 60,000:1. All of this impacts upon people whose web content is being accessed without them gaining any benefit, often at a cost to themselves.

However, I will end with what I think was the biggest news of the week, although as I said, it got hardly any coverage in the press. The Berlin Commissioner for Data Protection and Freedom of Information, announced in a press release that their office is calling on Google and Apple to remove the DeepSeek app after finding it in violation of the EU's General Data Protection Regulation (GDPR). Specifically they stated:

“The transfer of user data by DeepSeek to China is unlawful. DeepSeek has not been able to provide my office with convincing evidence that data of German users is protected in China at a level equivalent to that of the European Union. Chinese authorities have extensive access rights to personal data held by Chinese companies. In addition, DeepSeek users in China do not have enforceable rights and effective legal remedies as guaranteed in the European Union. I have therefore informed Google and Apple, as operators of the largest app platforms, of the violations and expect a prompt review of a blocking.”

This is MASSIVE 🧨💥💥💥💥

It will now be very difficult for other EU countries to continue to allow DeepSeek to access the data of EU citizens, and I suspect that the same will also apply to Manus and other new entrants in the AI-ecosphere from China. Given the lack of coverage, it may take some time for the effects of this to filter through, but make no mistake – this is as big a shockwave as when DeepSeek burst onto the scene at the start of 2025.

And who said tech stuff was boring!?

Stay informed. Stay critical. And wherever possible—stay ahead.

Regards

Tom Carter