Editor’s Pick

Enlarge / Helen Toner, former OpenAI board member, speaks during Vox Media’s 2023 Code Conference at The Ritz-Carlton, Laguna Niguel on September 27, 2023. (credit: Getty Images)

In a recent interview on “The Ted AI Show” podcast, former OpenAI board member Helen Toner said the OpenAI board was unaware of the existence of ChatGPT until they saw it on Twitter. She also revealed details about the company’s internal dynamics and the events surrounding CEO Sam Altman’s surprise firing and subsequent rehiring last November.

OpenAI released ChatGPT publicly on November 30, 2022, and its massive surprise popularity set OpenAI on a new trajectory, shifting focus from being an AI research lab to a more consumer-facing tech company.

“When ChatGPT came out in November 2022, the board was not informed in advance about that. We learned about ChatGPT on Twitter,” Toner said on the podcast.

Enlarge (credit: Flavio Coelho/Getty Images)

Two years ago when “Michael,” an owner of cryptocurrency, contacted Joe Grand to help recover access to about $2 million worth of bitcoin he stored in encrypted format on his computer, Grand turned him down.

Michael, who is based in Europe and asked to remain anonymous, stored the cryptocurrency in a password-protected digital wallet. He generated a password using the RoboForm password manager and stored that password in a file encrypted with a tool called TrueCrypt. At some point, that file got corrupted, and Michael lost access to the 20-character password he had generated to secure his 43.6 BTC (worth a total of about 4,000 euros, or $5,300, in 2013). Michael used the RoboForm password manager to generate the password but did not store it in his manager. He worried that someone would hack his computer and obtain the password.

“At [that] time, I was really paranoid with my security,” he laughs.

Enlarge (credit: Getty Images)

The US Treasury Department has sanctioned three Chinese nationals for their involvement in a VPN-powered botnet with more than 19 million residential IP addresses they rented out to cybercriminals to obfuscate their illegal activities, including COVID-19 aid scams and bomb threats.

The criminal enterprise, the Treasury Department said Tuesday, was a residential proxy service known as 911 S5. Such services provide a bank of IP addresses belonging to everyday home users for customers to route Internet connections through. When accessing a website or other Internet service, the connection appears to originate with the home user.

In 2022, researchers at the University of Sherbrooke profiled 911[.]re, a service that appears to be an earlier version of 911 S5. At the time, its infrastructure comprised 120,000 residential IP addresses. This pool was created using one of two free VPNs—MaskVPN and DewVPN—marketed to end users. Besides acting as a legitimate VPN, the software also operated as a botnet that covertly turned users’ devices into a proxy server. The complex structure was designed with the intent of making the botnet hard to reverse engineer.

Enlarge (credit: Getty Images)

On Monday, OpenAI announced the formation of a new “Safety and Security Committee” to oversee risk management for its projects and operations. The announcement comes as the company says it has “recently begun” training its next frontier model, which it expects to bring the company closer to its goal of achieving artificial general intelligence (AGI), though some critics say AGI is farther off than we might think. It also comes as a reaction to a terrible two weeks in the press for the company.

Whether the aforementioned new frontier model is intended to be GPT-5 or a step beyond that is currently unknown. In the AI industry, “frontier model” is a term for a new AI system designed to push the boundaries of current capabilities. And “AGI” refers to a hypothetical AI system with human-level abilities to perform novel, general tasks beyond its training data (unlike narrow AI, which is trained for specific tasks).

Meanwhile, the new Safety and Security Committee, led by OpenAI directors Bret Taylor (chair), Adam D’Angelo, Nicole Seligman, and Sam Altman (CEO), will be responsible for making recommendations about AI safety to the full company board of directors. In this case, “safety” partially means the usual “we won’t let the AI go rogue and take over the world,” but it also includes a broader set of “processes and safeguards” that the company spelled out in a May 21 safety update related to alignment research, protecting children, upholding election integrity, assessing societal impacts, and implementing security measures.

A previously unknown piece of ransomware, dubbed ShrinkLocker, encrypts victim data using the BitLocker feature built into the Windows operating system.

BitLocker is a full-volume encryptor that debuted in 2007 with the release of Windows Vista. Users employ it to encrypt entire hard drives to prevent people from reading or modifying data in the event they get physical access to the disk. Starting with the rollout of Windows 10, BitLocker by default has used the 128-bit and 256-bit XTS-AES encryption algorithm, giving the feature extra protection from attacks that rely on manipulating cipher text to cause predictable changes in plain text.

Recently, researchers from security firm Kaspersky found a threat actor using BitLocker to encrypt data on systems located in Mexico, Indonesia, and Jordan. The researchers named the new ransomware ShrinkLocker, both for its use of BitLocker and because it shrinks the size of each non-boot partition by 100 MB and splits the newly unallocated space into new primary partitions of the same size.

Enlarge / This is fine. (credit: Getty Images)

If you use Google regularly, you may have noticed the company’s new AI Overviews providing summarized answers to some of your questions in recent days. If you use social media regularly, you may have come across many examples of those AI Overviews being hilariously or even dangerously wrong.

Factual errors can pop up in existing LLM chatbots as well, of course. But the potential damage that can be caused by AI inaccuracy gets multiplied when those errors appear atop the ultra-valuable web real estate of the Google search results page.

“The examples we’ve seen are generally very uncommon queries and aren’t representative of most people’s experiences,” a Google spokesperson told Ars. “The vast majority of AI Overviews provide high quality information, with links to dig deeper on the web.”

Enlarge (credit: JAVS)

A software maker serving more than 10,000 courtrooms throughout the world hosted an application update containing a hidden backdoor that maintained persistent communication with a malicious website, researchers reported Thursday, in the latest episode of a supply-chain attack.

The software, known as the JAVS Viewer 8, is a component of the JAVS Suite 8, an application package courtrooms use to record, play back, and manage audio and video from proceedings. Its maker, Louisville, Kentucky-based Justice AV Solutions, says its products are used in more than 10,000 courtrooms throughout the US and 11 other countries. The company has been in business for 35 years.

JAVS Viewer users at high risk

Researchers from security firm Rapid7 reported that a version of the JAVS Viewer 8 available for download on javs.com contained a backdoor that gave an unknown threat actor persistent access to infected devices. The malicious download, planted inside an executable file that installs the JAVS Viewer version 8.3.7, was available no later than April 1, when a post on X (formerly Twitter) reported it. It’s unclear when the backdoored version was removed from the company’s download page. JAVS representatives didn’t immediately respond to questions sent by email.

Enlarge (credit: Getty Images)

Bing, Microsoft’s search engine platform, went down in the very early morning today. That meant that searches from Microsoft’s Edge browsers that had yet to change their default providers didn’t work. It also meant that services relying on Bing’s search API—Microsoft’s own Copilot, ChatGPT search, Yahoo, Ecosia, and DuckDuckGo—similarly failed.

Services were largely restored by the morning Eastern work hours, but the timing feels apt, concerning, or some combination of the two. Google, the consistently dominating search platform, just last week announced and debuted AI Overviews as a default addition to all searches. If you don’t want an AI response but still want to use Google, you can hunt down the new “Web” option in a menu, or you can, per Ernie Smith, tack “&udm=14” onto your search or use Smith’s own “Konami code” shortcut page.

If dismay about AI’s hallucinations, power draw, or pizza recipes concern you—along with perhaps broader Google issues involving privacy, tracking, news, SEO, or monopoly power—most of your other major options were brought down by a single API outage this morning. Moving past that kind of single point of vulnerability will take some work, both by the industry and by you, the person wondering if there’s a real alternative.

Enlarge

For more than four days, a server at the very core of the Internet’s domain name system was out of sync with its 12 root server peers due to an unexplained glitch that could have caused stability and security problems worldwide. This server, maintained by Internet carrier Cogent Communications, is one of the 13 root servers that provision the Internet’s root zone, which sits at the top of the hierarchical distributed database known as the domain name system, or DNS.

Here’s a simplified recap of the way the domain name system works and how root servers fit in:

When someone enters wikipedia.org in their browser, the servers handling the request first must translate the human-friendly domain name into an IP address. This is where the domain name system comes in. The first step in the DNS process is the browser queries the local stub resolver in the local operating system. The stub resolver forwards the query to a recursive resolver, which may be provided by the user’s ISP or a service such as 1.1.1.1 or 8.8.8.8 from Cloudflare and Google, respectively.

Enlarge / Nathan Benaich of Air Street Capital delivers the opening presentation on the state of AI at EmTech Digital 2024 on May 22, 2024. (credit: Benj Edwards)

CAMBRIDGE, Massachusetts—On Wednesday, AI enthusiasts and experts gathered to hear a series of presentations about the state of AI at EmTech Digital 2024 on the Massachusetts Institute of Technology’s campus. The event was hosted by the publication MIT Technology Review. The overall consensus is that generative AI is still in its very early stages—with policy, regulations, and social norms still being established—and its growth is likely to continue into the future.

I was there to check the event out. MIT is the birthplace of many tech innovations—including the first action-oriented computer video game—among others, so it felt fitting to hear talks about the latest tech craze in the same building that hosts MIT’s Media Lab on its sprawling and lush campus.

EmTech’s speakers included AI researchers, policy experts, critics, and company spokespeople. A corporate feel pervaded the event due to strategic sponsorships, but it was handled in a low-key way that matches the level-headed tech coverage coming out of MIT Technology Review. After each presentation, MIT Technology Review staff—such as Editor-in-Chief Mat Honan and Senior Reporter Melissa Heikkilä—did a brief sit-down interview with the speaker, pushing back on some points and emphasizing others. Then the speaker took a few audience questions if time allowed.