Welcome to The Tech Fix, a weekly newsletter where we break down the latest technology news and trends from the advertising and marketing industry, curated by technology editor Jessica Heygate.

Twitter’s identity under “free speech absolutist” Elon Musk became a lot clearer this week.

Tucker Carlson, the former Fox News host who was abruptly fired last month, reportedly over his extreme views, plans to make the platform his new home.

In a video posted on Tuesday, Carlson said that he was starting a “new version” of his former primetime Fox News show on Twitter that would champion free speech and provide an alternative to media organizations that are “full of lies.” “We’re back,” he declared.

Carlson did not provide details about the planned show, which may prompt Fox to pursue legal action. Under the terms of Carlson’s contract with Fox, he is prohibited from hosting a show on an alternate network.

Elon Musk responded that Twitter has “not signed a deal” with Carlson and that he will be subject to “the same rules and rewards” of any content creators on the platforms, including providing Carlson with a cut of advertising revenue.

Twitter will have a hard time finding large advertisers to support Carlson — the host lost the support of dozens of household brands such as Disney and T-Mobile in recent years over controversial statements made on his show about white supremacy and the Black Lives Matter movement, among other things.

I wonder what NBCUniversal, which last week extended its partnership with Twitter, thinks about the prospect of its Olympics coverage running in the same environment as Carlson’s political rants.

The news tumbled the stock price of Rumble, a conservative video network backed by billionaire Peter Thiel.

Google’s I/O developer conference on Wednesday was entirely dedicated to its work in AI, demonstrating how the technology is significantly evolving all of Google’s services.

One of the largest shifts is to Google’s search experience. Executives showcased a powerful, experimental version of search through which Google’s generative AI chatbot Bard can complete complex demands via integrations with a range of tools.

Google has integrated Gmail, Maps, its image recognition technology Lens and work tools Sheets and Docs into Bard, as well as a range of third-party applications including Adobe’s AI art generator Firefly, Spotify, Instacart, Walmart, Uber Eats and TripAdvisor.

The company dropped the waitlist for Bard, opening it up to 180 countries and territories, and added support for Japanese and Korean.

Ads within the new search experience look similar to how they are displayed today, in sponsored carousels and text results, but VP of engineering Cathy Edwards said the company was exploring different ways to integrate ads as search evolves.

Google’s presentation showcased a range of useful applications of AI, but the technology is equally dangerous. Critics include one of the masterminds of Google’s AI work, Geoffrey Hinton, who recently left the company after more than a decade in order to more freely speak about the dangers of the technology.

In a move seemingly designed to temper concerns, Google executives repeatedly said throughout the I/O presentation they were investing in AI “responsibly,” but they didn’t extrapolate much on what this means beyond tools to help users identify synthetically generated content. There was no indication that AI development was slowing down for guardrails to be established, as those concerned about AI have requested.

• A new digital solution from Scope3 automatically blocks ad inventory from platforms categorized as a high emissions contributor. The Climate Shield solution is available for activation across a range of DSPs.

• Supply-side platform PubMatic has launched a supply path optimization solution called Activate that offers more direct access to video inventory via private marketplace or programmatic guaranteed deals. Partners include Dentsu, FuboTV, GroupM, Havas Media Group, LG, Mars and Omnicom Media Group Germany.

• MiQ has launched a shopper intelligence platform, Consumer Link, which combines transaction and behavioral and TV data for CPG marketers to activate programmatic campaigns.

• Vericast has developed a contextual advertising solution that categorizes web content from more than 5 million new web pages daily and aggregates metadata from more than 18 million mobile apps. Vericast claims its algorithm scores websites within hours of publication.

• AdRoll has created an automation tool that allows marketers to create and analyze social ads, display ads, and email in a single interface.

• Code and Theory is working with Oracle Cloud Infrastructure to develop internal AI tools. The Stagwell company said it is leveraging Oracle “as a private cloud option to train and tune AI models to meet specific client business needs.” It claims this goes beyond the growing number of SaaS partnerships agencies have struck with firms such as OpenAI and Jasper.

• WhatsApp has expanded its payment feature to Singapore via a partnership with Stripe. Users will be able to pay businesses within the chat experience in Singapore, following trials in India and Brazil.

The FTC’s lawsuit against data broker Kochava was dismissed by a federal judge in Idaho last week, but the case is not yet over.

The judge ruled that the FTC lacked sufficient evidence to support its claims that Kochava’s data sales “puts consumers at significant risk.” In its August complaint, the FTC alleged that Kochava sold geolocation data from “hundreds of millions” of mobile devices that could be used to track people who visit “sensitive locations” including abortion clinics, places of worship, domestic violence shelters and addiction recovery facilities.

The regulator was granted 30 days to amend its arguments.

President Joe Biden urged the FTC to combat “digital surveillance” of reproductive care in a July 8 executive order, weeks after abortion rights in the U.S. were overturned. Shortly after, the FTC said that cracking down on “harmful commercial surveillance practices” and sensitive data sales is “a top priority.”

Financial Times technology reporter Cristina Criddle delivers a frightening account of how TikTok employees tracked her online activity via an account she had set up for her cat. Criddle claims that after publishing stories about TikTok’s toxic workplace culture, the company used her account to surveil her phone — including tracking her physical location — in a bid to figure out who her sources were. The surveillance of Criddle and another journalist was first revealed in December and is being investigated by the Justice Department. TikTok CEO Shou Zi Chew was grilled about the company’s surveillance tactics during the company’s congressional hearing in March.