Connect with us

Tech

Apple says researchers can vet its child safety features. But it’s suing a startup that does just that.

Published

on

Apple says researchers can vet its child safety features. But it’s suing a startup that does just that.


In 2019, Apple filed a lawsuit against Corellium, which lets security researchers cheaply and easily test mobile devices by emulating their software rather than requiring them to access the physical devices. The software, which also emulates Android devices, can be used to fix those problems.

In the lawsuit, Apple argued that Corellium violated its copyrights, enabled the sale of software exploits used for hacking, and shouldn’t exist. The startup countered by saying that its use of Apple’s code was a classic protected case of fair use. The judge has largely sided with Corellium so far. Part of the two-year case was settled just last week—days after news of the company’s CSAM technology became public. 

On Monday, Corellium announced a $15,000 grant for a program it is specifically promoting as a way to look at iPhones under a microscope and hold Apple accountable. On Tuesday, Apple filed an appeal continuing the lawsuit.

In an interview with MIT Technology Review, Corellium’s chief operating officer, Matt Tait, said that Federighi’s comments do not match reality.

“That’s a very cheap thing for Apple to say,” he says. “There is a lot of heavy lifting happening in that statement.”

“iOS is designed in a way that’s actually very difficult for people to do inspection of system services.”

“iOS is designed in a way that’s actually very difficult for people to do inspection of system services.”

Matt Tait, Corellium

He is not the only one disputing Apple’s position.

“Apple is exaggerating a researcher’s ability to examine the system as a whole,” says David Thiel, chief technology officer at Stanford’s Internet Observatory. Thiel, the author of a book called iOS Application Security, tweeted that the company spends heavily to prevent the same thing it claims is possible.

“It requires a convoluted system of high-value exploits, dubiously sourced binaries, and outdated devices,” he wrote. “Apple has spent vast sums specifically to prevent this and make such research difficult.”

Surveillance accountability

If you want to see exactly how Apple’s complex new tech works, you can’t simply look inside the operating system on the iPhone that you just bought at the store. The company’s “walled garden” approach to security has helped solve some fundamental problems, but it also means that the phone is designed to keep visitors out—whether they’re wanted or not. 

(Android phones, meanwhile, are fundamentally different. While iPhones are famously locked down, all you need to do to unlock an Android is plug in a USB device, install developer tools, and gain the top-level root access.) 

Apple’s approach means researchers are left locked in a never-ending battle with the company to try to gain the level of insight they require.

There are a few possible ways Apple and security researchers could verify that no government is weaponizing the company’s new child safety features, however.



Tech

Why detecting AI-generated text is so difficult (and what to do about it)

Published

on

Why detecting AI-generated text is so difficult (and what to do about it)


This tool is OpenAI’s response to the heat it’s gotten from educators, journalists, and others for launching ChatGPT without any ways to detect text it has generated. However, it is still very much a work in progress, and it is woefully unreliable. OpenAI says its AI text detector correctly identifies 26% of AI-written text as “likely AI-written.” 

While OpenAI clearly has a lot more work to do to refine its tool, there’s a limit to just how good it can make it. We’re extremely unlikely to ever get a tool that can spot AI-generated text with 100% certainty. It’s really hard to detect AI-generated text because the whole point of AI language models is to generate fluent and human-seeming text, and the model is mimicking text created by humans, says Muhammad Abdul-Mageed, a professor who oversees research in natural-language processing and machine learning at the University of British Columbia

We are in an arms race to build detection methods that can match the latest, most powerful models, Abdul-Mageed adds. New AI language models are more powerful and better at generating even more fluent language, which quickly makes our existing detection tool kit outdated. 

OpenAI built its detector by creating a whole new AI language model akin to ChatGPT that is specifically trained to detect outputs from models like itself. Although details are sparse, the company apparently trained the model with examples of AI-generated text and examples of human-generated text, and then asked it to spot the AI-generated text. We asked for more information, but OpenAI did not respond. 

Last month, I wrote about another method for detecting text generated by an AI: watermarks. These act as a sort of secret signal in AI-produced text that allows computer programs to detect it as such. 

Researchers at the University of Maryland have developed a neat way of applying watermarks to text generated by AI language models, and they have made it freely available. These watermarks would allow us to tell with almost complete certainty when AI-generated text has been used. 

The trouble is that this method requires AI companies to embed watermarking in their chatbots right from the start. OpenAI is developing these systems but has yet to roll them out in any of its products. Why the delay? One reason might be that it’s not always desirable to have AI-generated text watermarked. 

One of the most promising ways ChatGPT could be integrated into products is as a tool to help people write emails or as an enhanced spell-checker in a word processor. That’s not exactly cheating. But watermarking all AI-generated text would automatically flag these outputs and could lead to wrongful accusations.

Continue Reading

Tech

The original startup behind Stable Diffusion has launched a generative AI for video

Published

on

The original startup behind Stable Diffusion has launched a generative AI for video


Set up in 2018, Runway has been developing AI-powered video-editing software for several years. Its tools are used by TikTokers and YouTubers as well as mainstream movie and TV studios. The makers of The Late Show with Stephen Colbert used Runway software to edit the show’s graphics; the visual effects team behind the hit movie Everything Everywhere All at Once used the company’s tech to help create certain scenes.  

In 2021, Runway collaborated with researchers at the University of Munich to build the first version of Stable Diffusion. Stability AI, a UK-based startup, then stepped in to pay the computing costs required to train the model on much more data. In 2022, Stability AI took Stable Diffusion mainstream, transforming it from a research project into a global phenomenon. 

But the two companies no longer collaborate. Getty is now taking legal action against Stability AI—claiming that the company used Getty’s images, which appear in Stable Diffusion’s training data, without permission—and Runway is keen to keep its distance.

Gen-1 represents a new start for Runway. It follows a smattering of text-to-video models revealed late last year, including Make-a-Video from Meta and Phenaki from Google, both of which can generate very short video clips from scratch. It is also similar to Dreamix, a generative AI from Google revealed last week, which can create new videos from existing ones by applying specified styles. But at least judging from Runway’s demo reel, Gen-1 appears to be a step up in video quality. Because it transforms existing footage, it can also produce much longer videos than most previous models. (The company says it will post technical details about Gen-1 on its website in the next few days.)   

Unlike Meta and Google, Runway has built its model with customers in mind. “This is one of the first models to be developed really closely with a community of video makers,” says Valenzuela. “It comes with years of insight about how filmmakers and VFX editors actually work on post-production.”

Gen-1, which runs on the cloud via Runway’s website, is being made available to a handful of invited users today and will be launched to everyone on the waitlist in a few weeks.

Last year’s explosion in generative AI was fueled by the millions of people who got their hands on powerful creative tools for the first time and shared what they made with them. Valenzuela hopes that putting Gen-1 into the hands of creative professionals will soon have a similar impact on video.

“We’re really close to having full feature films being generated,” he says. “We’re close to a place where most of the content you’ll see online will be generated.”

Continue Reading

Tech

When my dad was sick, I started Googling grief. Then I couldn’t escape it.

Published

on

The Download: trapped by grief algorithms, and image AI privacy issues


I am a mostly visual thinker, and thoughts pose as scenes in the theater of my mind. When my many supportive family members, friends, and colleagues asked how I was doing, I’d see myself on a cliff, transfixed by an omniscient fog just past its edge. I’m there on the brink, with my parents and sisters, searching for a way down. In the scene, there is no sound or urgency and I am waiting for it to swallow me. I’m searching for shapes and navigational clues, but it’s so huge and gray and boundless. 

I wanted to take that fog and put it under a microscope. I started Googling the stages of grief, and books and academic research about loss, from the app on my iPhone, perusing personal disaster while I waited for coffee or watched Netflix. How will it feel? How will I manage it?

I started, intentionally and unintentionally, consuming people’s experiences of grief and tragedy through Instagram videos, various newsfeeds, and Twitter testimonials. It was as if the internet secretly teamed up with my compulsions and started indulging my own worst fantasies; the algorithms were a sort of priest, offering confession and communion. 

Yet with every search and click, I inadvertently created a sticky web of digital grief. Ultimately, it would prove nearly impossible to untangle myself. My mournful digital life was preserved in amber by the pernicious personalized algorithms that had deftly observed my mental preoccupations and offered me ever more cancer and loss. 

I got out—eventually. But why is it so hard to unsubscribe from and opt out of content that we don’t want, even when it’s harmful to us? 

I’m well aware of the power of algorithms—I’ve written about the mental-health impact of Instagram filters, the polarizing effect of Big Tech’s infatuation with engagement, and the strange ways that advertisers target specific audiences. But in my haze of panic and searching, I initially felt that my algorithms were a force for good. (Yes, I’m calling them “my” algorithms, because while I realize the code is uniform, the output is so intensely personal that they feel like mine.) They seemed to be working with me, helping me find stories of people managing tragedy, making me feel less alone and more capable. 

In my haze of panic and searching, I initially felt that my algorithms were a force for good. They seemed to be working with me, making me feel less alone and more capable. 

In reality, I was intimately and intensely experiencing the effects of an advertising-driven internet, which Ethan Zuckerman, the renowned internet ethicist and professor of public policy, information, and communication at the University of Massachusetts at Amherst, famously called “the Internet’s Original Sin” in a 2014 Atlantic piece. In the story, he explained the advertising model that brings revenue to content sites that are most equipped to target the right audience at the right time and at scale. This, of course, requires “moving deeper into the world of surveillance,” he wrote. This incentive structure is now known as “surveillance capitalism.” 

Understanding how exactly to maximize the engagement of each user on a platform is the formula for revenue, and it’s the foundation for the current economic model of the web. 

Continue Reading

Copyright © 2021 Seminole Press.