People are already using chatbots as therapists, with the emergence of generative AI raising new questions around tech's role in mental health, Axios' Peter Allen Clark reports.

Why it matters: Virtually no one is suggesting you replace a compassionate human professional with a probability-driven neural network — but plenty of people seeking info or help say they appreciate the approachability (and low cost) of an onscreen text box.

What’s happening: Users are filling online forums with accounts of their experiences casting ChatGPT as their personal therapist.

• In the ChatGPT subreddit, it's easy to find people offering examples of addressing trauma or attempting to improve communication skills with the tech.
• Others are sharing advice on what kind of prompts to use and how to get the best responses in a ChatGPT therapy session.
• The low cost isn't the only lure — users also praise the accessibility of the tech and the comfort they feel in engaging with it.

What they're saying: "As someone who has consumed a lot of mental health services in his life, I can say that I found [ChatGPT] to be incredibly helpful, much more than many of the humans I have interacted with," one Reddit user shared.

The other side: ChatGPT creator OpenAI's policies say its tech is not to be used to tell "someone that they have or do not have a certain health condition, or providing instructions on how to cure or treat a health condition."

• "OpenAI's models are not fine-tuned to provide medical information," the policies say.

The big picture: Platforms that offer mental health services through text, like BetterHelp, have flourished in the pandemic era. And a growing number of them are specifically offering a chatbot.

• Recent apps like Wysa, Limbic and Replika all offer users AI-driven conversations about mental health. While some cast themselves as a companion to talk therapy, others, like Replika, offer a "companion" who is "always ready to chat when you need an empathetic friend."
• These apps' popularity has also raised alarms over their effectiveness and their ability to protect users' privacy.

Meanwhile, many mental health professionals are cautioning users against replacing the personal approach of therapy with a chatbot.

• "Only a therapist can provide a personalized or customized treatment plan for you, and that takes some time and that gets actualized as you are making progress," therapist Daniela Marin said in a YouTube video. "It won't keep you accountable, it doesn't care if you do or don’t do the work."
• Licensed marriage and family therapist Emma McAdam said recently that ChatGPT "can never provide a supportive relationship and the motivational supportive structure of actual therapy with a real person."

Between the lines: Generative AIs like ChatGPT have trouble distinguishing between fact and fiction, and at one point Microsoft's Bing chatbot seemed to be displaying mental disorders of its own.

• For all the excitement, ChatGPT today is, in the words of OpenAI CEO Sam Altman, "a horrible product," and users who turn to it for therapeutic help are proceeding at their own risk.

TikTok's access controls on U.S. user data are much weaker than the company says, a former ByteDance employee told the office of Sen. Josh Hawley (R-Mo.), per a letter from Hawley to Treasury Secretary Janet Yellen shared exclusively with Axios' Ashley Gold.

Driving the news: The whistleblower's allegations, which have not been independently seen or verified by Axios, suggest that TikTok overstates its separation from its China-based owner ByteDance, relies on proprietary Chinese software that could have back doors, and uses tools that allow employees to easily toggle between U.S. and Chinese user data.

What they're saying: "This whistleblower's allegations are deeply concerning. They also appear to contradict public statements made by TikTok and ByteDance executives," Hawley writes.

• He cites Congressional testimony from last September by TikTok chief operating officer Vanessa Pappas stating "there are strict access controls around the data that is accessed in the United States," along with reporting from Forbes and Reuters about Bytedance employees improperly accessing TikTok's U.S. user data.
• The whistleblower told Hawley's office, per the letter: "I have seen first-hand China-based engineers flipping over to non-China datasets and creating scheduled tasks to backup, aggregate, and analyze data," adding that "TikTok and ByteDance are functionally the same company."

Context: Calls to ban or limit TikTok and other Chinese companies are popular in Congress right now. Meanwhile, a review of the company's practices by the Committee on Foreign Investment in the United States (CFIUS), meant to ultimately result in a security deal, has dragged on for more than two years.

• TikTok has consistently maintained it keeps U.S. user data secure and that it is committed to a strong CFIUS deal.

Read the full letter.

The FTC has asked Twitter for information about recent layoffs as well as details on the firm's sharing of internal documents with journalists. Meanwhile, the EU wants Twitter to hire more content moderators, according to the Financial Times.

Why it matters: Regulators around the globe have warned Twitter that, however dire its need to cut costs, it must still meet legal obligations.

Driving the news: The Wall Street Journal reported that the FTC has sent more than a dozen letters to Twitter since Elon Musk took over. Among the concerns is whether recent layoffs could imperil the company's ability to adhere to the terms of a consent decree. The agency is also seeking to interview Musk, the WSJ and the New York Times reported.

• The Financial Times reports, citing sources, that the EU has told Twitter it needs to hire more content moderators to augment the work being done by AI and volunteer fact-checkers.
• Musk himself also made headlines, both for telling investors at a Morgan Stanley conference that Twitter's finances are improving as well as for insulting — and then later apologizing to — a former employee with a disability.

Between the lines: The FTC's inquiry into Twitter has already become a political lightning rod.

• The Judiciary Committee issued a press release blasting the FTCs move: "There is no logical reason, for example, why the FTC needs to know the identities of journalists engaging with Twitter. "

Be smart: Twitter is under an FTC consent decree requiring it to protect customer information that could have been exposed when Musk opened Twitter's systems to several writers for the "Twitter Files."

• "Protecting consumers' privacy is exactly what the FTC is supposed to do," FTC spokesman Douglas Farrar said. "It should come as no surprise that career staff at the commission are conducting a rigorous investigation into Twitter's compliance with a consent order that came into effect long before Mr. Musk purchased the company."

On Tap

• Spotify is hosting a Stream On media event at 10am PT.
• The Senate Homeland Security Committee is having a hearing on the risks and opportunities of AI.
• Asana and MongoDB report quarterly earnings after the markets close.

Trading Places

• Gigi Sohn has withdrawn from consideration to fill the vacant fifth commissioner spot at the FCC after a prolonged, 16-month effort win Senate confirmation.

ICYMI

• Google has scheduled its I/O developer conference for May 10 in Mountain View.
• Apple added a yellow version of the iPhone 14 (and iPhone 14 Plus), continuing a recent tradition of introducing an additional color in the middle of its model year. (The Verge)

I've heard of cars crashing into checkstops. But in this case it was Czech Stop, a beloved bakery in the Texas city of West.