Google now uses your uploaded search media to train AI, DeepSeek may be developing its own AI chips, and Meta is facing $1.4 trillion in state lawsuits

**SPEAKER_1** (0:00)
Heat up your 4th of July at the Home Depot with our wide variety of grills under $300 and make every gathering one to remember. Give your outdoor space a glow up. Whatever your budget is, the savings on seasonal plants starting at $5.
With the grill fired up and your backyard set to perfection, you'll be able to invite friends and family over to kick off the party. Start celebrating with low prices guaranteed at the Home Depot. Prices may vary by store, using the pricey Home depot.com/pricematch for details.

**SPEAKER_2** (0:37)
Google is now using your uploaded search media to train AI.
DeepSeek may be developing its own AI chips.
And Meta is facing $1.4 trillion in state lawsuits over social media addiction. It's Tuesday, July 7th. And here's a quick look at some of the news happening in and around tech this morning from Engadget.
Well, Google is at it again. The company recently, and quietly, introduced a change to how it hoovers up your data to train its AI platforms. It can now scoop up media that you upload to its various search tools for training purposes. That's according to a report by TechCrunch. Now, this includes images, files, and audio and video recordings. Well, that's pretty much everything, right? You know, what does this mean exactly? Well, if you upload a photo to Google Lens to search for something visually, the company can take it. The same goes for the audio accompanying any Google voice search and anything uploaded to Google Translate. This applies to all search-related products. So, stuff like your personal Google photos are safe for now. Every user is automatically opted in as the gaping mall of generative AI needs data to feed on and it's running out. There is, however, a solution for those who do not like random mega-corporations pouring through their images and videos. You can opt out of this by changing some particular settings. You can head to the dedicated Search Services history page and uncheck the Save Media box. You know, you can always turn off AI overview results by popping in the, you know, dash AI before a query. Of course, you know, this is modern AI. It doesn't always need permission to get our data, right? Yeah, just ask some musicians.
DeepSeek, the Chinese AI company, which sometimes keeps tech executives awake at night, is reportedly looking to build its own silicon. Reuters sources suggest that DeepSeek is looking to build a chip for inference, which is the process of running a model that has already been trained and built. The move comes as part of a push to reduce its reliance on third-party chip providers like Huawei and NVIDIA. If the report is accurate, it could signal another entrant in the fiercely competitive Chinese AI market. The report suggests that DeepSeek is already in talks with manufacturing partners and has quietly begun hiring engineers to support the effort. Of course, given the various import and export bans, we probably will not see any of DeepSeek's tech leaving China. But given how the company made its name with an open-source model that could hold its own against the big names for a fraction of the cost, its interest in chip development should be taken seriously. Especially if it finds a way to create similar cost and power savings, which would likely put another dent in NVIDIA's stock.
Meta is facing penalties of up to a massive $1.4 trillion from four US states that sued the company over the addictive designs of Facebook and Instagram. That's according to a Reuters report. Those states, California, Colorado, Kentucky, and New Jersey, also accused the company of misleading the public about the safety of those apps. The figure is close to Meta's $1.5 trillion dollar market capitalization. At a court hearing last month, the states said they calculated the penalties by estimating the number of young users affected by Meta's platforms and multiplying that by fines set by state law. Meta revealed the figure in response to a request from the state's attorneys general on how penalties should be calculated, but said that the amount was unjustified. Mark Zuckerberg's company is facing additional lawsuits from 49 other states, not part of the $1.4 trillion suit. Most of those allege that the company violated the Federal Children's Online Privacy Protection Act by collecting data from kids without the required parental consent. US. District Judge Yvonne Gonzalez-Rogers will address those claims, plus the four state lawsuits at a trial in August. Another 14 states have brought claims based on local laws that will be heard at a separate trial next February.
Meta has previously denied the allegations on the basis that social media addiction is not an established psychiatric condition. The company's Instagram head Adam Masseri previously compared it to being addicted to a Netflix show. In response, the American Psychiatric Association said that social media addiction is not currently listed as a diagnosis in the Diagnostic Manual, but that does not mean it does not exist. Juries have found merit to state's claims before, with one recently awarding New Mexico $375 million after finding that Meta misled consumers in the state. Meta and other social networks also recently paid $27 million to settle a Kentucky school district lawsuit over similar claims.