Using ChatGPT as a therapist? Maybe not so safe, here’s why!

My friend recently texted me, “ChatGPT actually gets it,” referring to how they have been using the AI for emotional support. That got me thinking, many people are turning to AI for way more than just fact-checking. It’s a brilliant use of tech, but safety concerns? Well, OpenAI’s CEO revealed on a podcast this week that your “confidential” chats with ChatGPT might not be so private after all. Like I did the moment I heard, you’ll want to share this with your Gen Z friends who treat ChatGPT like their go-to confidant. Anyway, we’ve got a lot more AI news from this week, so enjoy!

OpenAI CEO, Sam Altman recently cautioned users about the lack of legal confidentiality when using ChatGPT for sensitive conversations, such as seeking therapy or life advice. Speaking on This Past Weekend w/ Theo Von podcast, Altman noted that, unlike traditional doctor-patient or attorney-client privilege, there’s no legal framework protecting personal discussions with AI.

He highlighted that many users, particularly younger ones, turn to ChatGPT for guidance on personal matters. However, in legal scenarios, OpenAI could be required to disclose these conversations, posing a privacy risk. This issue is evident in a current lawsuit with The New York Times, where a court has ordered OpenAI to preserve user chats (excluding ChatGPT Enterprise users). OpenAI is appealing, arguing the order oversteps, but it underscores the broader challenge of AI data privacy.

With growing concerns about digital privacy, Sam stressed the need for legal protections similar to those for therapists or doctors to ensure user trust and wider AI adoption. Until such safeguards are in place, you should exercise caution when sharing sensitive information with ChatGPT.

Google has entered the growing field of AI-driven “vibe-coding” with Opal, a new tool now being tested in the U.S. through Google Labs. Designed to make app creation accessible, Opal allows users to build mini web applications using simple text prompts, no coding expertise required. With Opal, users describe the app they envision, and Google’s AI models generate it. A visual editor displays the workflow, input, output, and generation steps, allowing users to refine prompts or add steps manually via a toolbar. Once complete, apps can be published online and shared for others to test using their Google accounts.

Unlike Google’s AI Studio, which targets developers, Opal’s intuitive interface aims to attract a broader audience, including non-technical users. This development aligns Google with competitors like Canva, Figma, and Replit, all working to empower everyone to create app prototypes without coding.

Spotify has partnered with Netflix to launch a mini game within its app, celebrating the release of Happy Gilmore 2, the sequel to Adam Sandler’s 1996 comedy, now streaming on Netflix as of July 25. This marks Spotify’s first collaboration with another company to create a gaming experience, signaling a new direction for the music platform.

Developed with MiK Studio, the Happy Gilmore 2 Tournament game invites users to perfect their virtual golf swing by tapping their device to complete three holes. The experience is enriched with audio clips from the film, a personalized “happy” playlist curated for each player, and fun Easter eggs like Happy’s boots, hockey stick, and an alligator. After playing, users can watch the movie’s trailer and share their scores and playlists on social media. While Spotify previously introduced a hidden game, Eat This Playlist, in 2023, this collaboration with Netflix is a notable step into gaming to boost user engagement, leveraging the nostalgia of a cult classic. The initiative could serve as a test to gauge interest in future gaming features on the platform.

YouTube is enhancing its Shorts platform with innovative generative AI features, including an image-to-video tool and new AI effects, announced on July 23. These tools aim to empower creators to produce dynamic, engaging content with ease.

The image-to-video feature allows users to transform a single photo from their camera roll into a six-second video. By uploading an image, creators receive tailored suggestions to animate it, whether adding motion to landscapes, bringing everyday photos to life, or animating group shots. For example, a static pedestrian signal image could become a short video zooming into a dancing figure. This tool, powered by Google’s Veo 2 AI model, is rolling out over the next week in the United States, Canada, Australia, and New Zealand, with plans for broader availability later this year.

Also, YouTube is introducing AI effects that let creators turn doodles into artistic visuals or transform selfies into imaginative videos, such as swimming underwater or appearing alongside a twin. These effects are accessible via the “Effects” icon in the Shorts camera under the “AI” tab. YouTube also launched AI Playground, a hub for generative AI tools, inspirational examples, and prefilled prompts, available by tapping the Create button and sparkle icon. All AI-generated content includes SynthID watermarks and clear labels for transparency.

Moby AI delivers accurate ecommerce analytics, real-time forecasting, and automated insights. Build AI analytics reports and make data-driven decisions.