Weekly Roundup: September 3-9, 2025
Regulatory action continues, power concerns are on the rise, and Sam Altman is apparently just like us
Regulators, mount up
China’s social media platforms rush to abide by AI-generated content labelling law
The Google antitrust ruling gives its AI rivals one big reason to cheer
FTC prepares to grill AI companies over impact on children, WSJ reports
Regulation and legal action were loaded topics in the AI space last week. In China, a new law requiring AI generated content to be labeled for the public. This includes explicit and implicit labels that will create cues in the interface and metadata. The law goes farther, also placing the onus of identifying violators and enforcing the labels for companies that house AI generated content. In the US Google scored a big win in its anti-trust case… for… now? It was determined that Google has, in fact, operated as an illegal monopoly in internet search. However, for the moment the only action to be taken is that it must share its search index with competitors. The market responded, as Google stock jumped 8%. Finally, the FTC is preparing to investigate the negative impacts of AI tools on mental health, with a specific focus on children. This in the wake of both teen deaths and allowing children to access inappropriate chatbots online. OpenAi, Meta, and Character.ai appear to be the main targets.
Gut check: this is going to get really messy as public concern around AI forces action. We will continue to see these kinds of cases, investigations, and rulings that will begin to build out the nearly non-existent case law around AI. New lawyers - pick this as your speciality. China is only a few steps behind the EU in terms of regulations, and there are other countries like France and Spain looking to pass their own laws. In the US… it seems to be mostly performative at the federal level, but I predict local lawmakers will follow Ohio and start drafting legislation. And while child protection is a great inroads for mental health, it will be important to watch the FTC investigation to get a sense of what will happen here for adults.
AI’s Infrastructure traffic jam
AI’s infrastructure problem is bigger than we think—here’s how to solve it
CenterPoint braces for rising power demand from Foxconn’s Houston AI projects with Apple and Nvidia
Data Centers That Don’t Exist Yet Are Already Haunting the Grid
As the demand for AI grows, it appears that the infrastructure that enables it is not matching pace. The most concerning is that of power for the expanding data centers being built all over the world. In places like the US where power grids are already being overburdened, this could cause very serious issues for residents where these data centers are built. The paradox is that companies are already planning for data centers that may never exist. This means that power companies are getting hit from all sides. They are having to invest in a growing need while also attempting to update a failing grid.
Gut check: Texas already went through a major power grid crisis in 2021, and this was without growth in data centers in areas like Houston. Consumers across the US get alerts asking them to reduce power consumption in advance of heat waves or cold snaps - they find that annoying enough, what happens when the ask is about AI? Not only are we dealing with an infrastructure issue, but a public perception issue as well. Will this force regulation, more responsible power usage by tech, investment in the grid from AI companies? Only time will tell.
Even Sam Altman thinks AI bots are annoying
Gut check: what is a metaphorical hot dog suit? Why is he wearing it metaphorically? This headline is so unhinged. All I can say is we feel the same, Sam, same.
