Kalakshar US

A week after being outed for stealthily using AI to produce high school sports reports and publicly “pausing” the project, mega-publisher Gannett has reportedly had to recheck each and every post the AI had written. Did we really learn nothing from CNET’s ignoble AI escapades in January?  

Gannett operates a number of regional and national publications including USA Today, The Arizona Republic and The Detroit Free Press. The company devised its “Lede AI” as a means of automating the droll work of summarizing the box scores of local high school sports leagues — a task the AI proved wholly incapable of. One such article read:

The Hardin County Tigers defeated the Memphis Business Execs 48-12 in a Tennessee high school football game on Friday. Hardin County scored early and often to roll over Memphis Business 48-12 in a Tennessee high school football matchup.

“High school reporting is different from covering college or professional sports,” On anonymous Gannett sports writer told Yahoo News. “And high school reporting can go underappreciated, but it’s extremely important. You’re covering a community.”  

“You’re not writing for as big of an audience, but you’re writing for a very, very specific one,” they added. “Family members — uncles, parents, people who care that your story has their kids’ names. They’re looking for keepsakes, things they can remember from their kids’ high school career.”

In response to the criticism, Gannett has elected to “pause” its use of the AI for the time being though the company made no mention of abandoning its use entirely. The company has also reportedly rechecked and updated every AI-written post for factual accuracy. The blurb above now simply reads: “The Hardin County Tigers defeated the Memphis Business Execs 48-12 in a Tennessee high school football game on Friday.“

Gannett’s dalliance with AI writers is the latest in a growing industry trend that has seen Buzzfeed shutter its newsroom entirely in favor of AI-generated content, Google develop a genAI for writing news posts and Apple build a foundational rival to ChatGPT. Even the AP is adapting its approach to accommodate the new technology.    

This article originally appeared on Engadget at https://www.engadget.com/usa-todays-publisher-had-to-update-all-of-the-sports-posts-its-ai-reporter-botched-215915908.html?src=rss 

The home state of some of the most influential AI companies has a new plan to confront the potential regulation of generative AI. California Governor Gavin Newsom signed an executive order instructing agencies in the state to study potential risks and use cases for the technology.

Under the order, state agencies are tasked with identifying “the most significant and beneficial uses of GenAI in the state” and creating frameworks to train state employees on how to use “state-approved” generative AI tools in their work. Likewise, it directs the same agencies to analyze potential negative impacts of the technology, including its effect on vulnerable communities and threats to “critical energy infrastructure” in the state.

The order also lays the groundwork for new partnerships with University of California at Berkeley and Stanford University, which will help study how generative AI is affecting the state’s workers. “This is a potentially transformative technology – comparable to the advent of the internet – and we’re only scratching the surface of understanding what GenAI is capable of,” Newsom said in a statement. “We recognize both the potential benefits and risks these tools enable.”

As Bloombergpoints out, the California order comes ahead of a White House executive order on generative AI expected in the coming months. Leaders of several top AI companies have been meeting with White House officials over the last severalmonths as the administration weighs potential regulations for AI. More than half a dozen companies have already agreed to some safeguards, including new investments in cybersecurity, following those meetings.

This article originally appeared on Engadget at https://www.engadget.com/california-governor-gavin-newsom-signs-executive-order-to-study-generative-ai-222733401.html?src=rss 

Google Workspace users are getting a new feature to prevent colleagues from messing with their documents. Users will now have the ability to lock their Google Drive files, which should help reduce unwanted commentary and changes when sharing with multiple team members. While this isn’t the most revolutionary feature, it could certainly come in handy — especially for companies that use and share a lot of publicly viewable documents.

Locking a document puts it in read-only mode, meaning viewers can’t edit, comment or leave suggestions on it. The document will remain in this read-only mode until unlocked by its owner. In a blog post, Google stated that this new feature will “streamline the process of restricting files in Drive.”

Previously, you could lock down a document or file when sharing it, or through a Google Drive API call. The new lock feature simplifies things: Right-click on the file you’d like to lock, scroll to file information from the popup menu, and select lock.

The file-locking feature has already been rolled out to some Rapid Release domains and will continue over the next 15 days. Folks on Scheduled Release domains can expect to start seeing the feature on September 20th, with a similar 15-day rollout period.

This article originally appeared on Engadget at https://www.engadget.com/google-workspaces-file-lock-stops-colleagues-messing-with-your-documents-194625235.html?src=rss 

The non-profit Mozilla Foundation deemed cars the “worst product category” ever reviewed for data privacy, according to research released Wednesday. Its Privacy Not Included Research division reviews everything from smart home devices to health and wellness apps. But of the 25 car brands the research team studied, not a single one passed the reviews, with top brands like Tesla, Nissan and Hyundai landing at the top of the worst-of-the-worst list. 

Tesla earned a mark against it for untrustworthy use of AI, making it the second product reviewed by Mozilla to fail every privacy test conducted. That’s because its AI-powered autopilot feature caused several deaths and hundreds of crashes. Meanwhile, companies like Nissan and Kia say they can collect information about your sexual activity and sex life, and Hyundai promises to comply with “lawful requests, whether formal or informal” to share your information with government and law enforcement. 

That left Renault, Dacia and BMW as the “least creepy” car options. Researchers couldn’t confirm whether Renault, which also owns Dacia, encrypts the data it collects, and it doesn’t go much beyond what’s required by data privacy law, but compared to others it’s not the worst. The reason for BMW landing higher on the list was also marginal, as researchers based it on the fact that the car maker doesn’t explicitly say they sell data to third parties for advertising purposes, while other manufacturers explicitly claim to do so. “From our reading of BMW’s privacy policy, they might not do this. But we’re also not 100% sure they don’t,” the researchers wrote. 

Still, every brand of car collected too much personal data, and most of them share or sell that information to a third party. The researchers spent 600 hours analyzing privacy policies, investigating app features and working directly with the car companies themselves to determine privacy rankings, but still concluded it was one of the more confusing categories they’ve tested. 

“Sorting through the large and confusing ecosystem of privacy policies for cars, car apps, car connected services, and more isn’t something most people have the time or experience to do,” members of Mozilla’s Privacy Not Included team wrote in a blog post. That leaves little for car buyers to do if they’re looking for an option that takes data privacy seriously because, at least according to Mozilla, they really are all that bad. 

This article originally appeared on Engadget at https://www.engadget.com/tesla-tops-mozillas-list-of-creepiest-carmakers-but-25-brands-failed-basic-data-privacy-tests-202017058.html?src=rss