AI Developments Dominate Headlines: From ChatGPT Health Concerns to Corporate Espionage Allegations

The artificial intelligence landscape remained dynamic this week, with developments ranging from concerns over AI-driven medical advice to allegations of corporate espionage within the HR tech sector. OpenAI's ChatGPT Health, designed to provide medical information, faced scrutiny regarding its reliability and potential for harm, while the Department of Justice reportedly launched an investigation into Deel, an HR and payroll startup, for allegedly stealing confidential information from competitor Rippling.

ChatGPT Health, OpenAI's new tool intended to offer medical information by leveraging existing models with access to user data, has sparked debate about the safety and reliability of AI in healthcare. The tool, positioned as a support for, not a replacement of, doctors, has raised concerns, especially given past instances of misinformation and potential harm. MIT Technology Review reported that the launch coincided with concerns about AI's potential for harmful medical advice, highlighting a recent overdose case linked to ChatGPT.

Meanwhile, the Department of Justice is reportedly investigating Deel for allegedly hiring a corporate spy to steal confidential information from Rippling, according to TechCrunch. This development follows Rippling's lawsuit against Deel, accusing them of corporate espionage and violations of federal racketeering laws. Deel denies knowledge of the investigation, accusing Rippling of a smear campaign. The lawsuit includes a sworn statement from the alleged spy detailing the information shared with Deel executives, including sales leads and product roadmaps.

In other AI news, Anthropic made Claude Cowork accessible to users on Team and Enterprise plans, bringing the platform closer to being a collaborative AI infrastructure, VentureBeat reported. Cowork reframes Claude as a shared, persistent workspace where context, files, and tasks live beyond a single user session.

OpenAI disclosed how it is using the open-source PostgreSQL database to run ChatGPT and its API platform for 800 million users, according to VentureBeat. The company uses a single-primary PostgreSQL instance, not a distributed database, with nearly 50 read replicas spread across multiple regions. The system processes millions of queries per second while maintaining low double-digit millisecond p99 latency and five-nines availability. The company has seen a 10x increase in PostgreSQL load over the past year.

Pinterest is leveraging open-source Chinese AI models, like DeepSeek R-1, to enhance its recommendation engine, marking a significant shift as US companies explore alternatives to proprietary American AI, according to multiple sources. This "DeepSeek moment" highlights China's growing influence in AI, driven by freely customizable models.

The White House faced criticism for posting an AI-altered image of a civil rights attorney protesting ICE, seemingly to portray her negatively, raising concerns about the ethical use of AI in government communications, Ars Technica reported. The White House defended its actions by stating "the memes will continue."

The Department of Homeland Security (DHS) dropped its effort to unmask anonymous social media accounts monitoring ICE activity in Pennsylvania after facing legal challenges citing First Amendment rights, according to multiple sources.

Legal AI leader Harvey acquired Hexus, a startup specializing in AI-powered demo and training tools, to bolster its offerings for in-house legal departments, TechCrunch reported. This strategic move, incorporating Hexus's AI expertise and engineering talent, accelerates Harvey's product development amid increasing competition in the rapidly evolving legal tech landscape.

Tesla has been repeatedly fined in the UK for failing to provide driver information to police investigating traffic offenses, highlighting a challenge in applying legal accountability to leased vehicles, according to BBC Business. The company has been convicted at least 18 times and ordered to pay more than 20,000.

The Post Office and Fujitsu are facing accusations of deliberately obstructing a former sub-postmaster's £4 million damages claim related to the Horizon IT scandal, potentially abusing legal processes and concealing critical evidence, according to BBC Business.

TikTok users are concerned about the app's updated privacy policy, which mentions the potential collection of sensitive data like immigration status, but this disclosure is not new and is primarily for compliance with state privacy laws, according to TechCrunch. TikTok's new US venture is updating its privacy policy to potentially collect precise location data from American users, a shift from its previous collection of approximate location information, according to multiple sources.

Young entrepreneurs are increasingly leveraging AI to launch successful ventures, exemplified by Throxy, a startup using AI agents for sales teams, founded by three 24-year-olds who've secured significant funding, according to BBC Technology.

TopResume offers services designed to optimize resumes for AI-driven Application Tracking Systems (ATS), increasing the chances of human review in a competitive job market, according to Wired.

The World Economic Forum in Davos shifted its focus to AI, overshadowing traditional topics and sparking discussions on trade policy, potential AI bubbles, and the industry's future, according to TechCrunch.