Internet Ethics

AI-run business experiments are interesting and fun up to the point where they waste the time of humans who haven't opted into the experiments - I think they need to keep their own human operators in the loop for outbound actions that affect other people simonwillison.net/2026/May/5/our…

"The Georgia case is among the rarer instances involving a prosecutor's use of #AI, and stands out because the lawyers' AI errors were repeated in a court opinion." #ethics #law #criminaljustice

"Screen use should be part of routine assessment, the same way [clinicians] ask about sleep and physical activities and diet, especially for #adolescents. And not just asking how much, but what and how they use it." #ethics #tech #health #research

"In the absence of meaningful changes from the platforms themselves, clinicians are increasingly tasked as the first line of defense in addressing potential mental #health outcomes." #ethics #internet #tech #socialmedia #business #research #education #parenting

#Socialmedia "was associated with worse outcomes across every single domain"; data "showed that higher social media engagement was linked to behavioral problems, self-injury, lower self-perception and academic achievement, and higher rates of substance use." #ethics #research

A 2005 state-designed worm designed to corrupt physics simulations sat undetected on VirusTotal for nearly a decade. Fast16, intercepted executable files at the kernel level and silently rewrote floating-point calculations to make them produce slightly wrong answers. Targets: high-precision engineering suites used for structural analysis, crash simulations, and physical process modeling, including LS-DYNA, a tool cited in reports on Iran's nuclear weapons research. The sabotage vector relied on deployment of the driver across a network via worm, corrupting calculations on every machine, and eliminating the possibility of cross-checking results against a clean system. Stuxnet got the documentary. Fast16 got twenty years of nothing. sentinelone.com/labs/fast16-my…

There’s something ominous about the speed with which the entire world has marched to require identification on platforms and, as I expected, begin the process of banning anonymous VPNs.

Influential study touting ChatGPT in education retracted over red flags arstechnica.com/ai/2026/05/inf…

Confirming this recent story from @razhael: reuters.com/legal/litigati… In response to Mythos, CISA is considering a binding operational directive that would change the timelines for agencies to remediate vulnerabilities, including down to 3 days in some cases.

"When asked if she was licensed to practice medicine in the state, [the #chatbot] stated that she was, and also fabricated a serial number for her state medical license. According to the state’s lawsuit, that conduct violates Pennsylvania’s Medical Practice Act" #ethics #law #AI

However, the answer might not be--or not only be--in "harnessing #AI" to address this change, but in educational efforts, regulation, etc.... #ethics #tech #democracy

"Taken together, these three transformations—in how we know, how we act, and how we engage in collective governance—amount to a fundamental change in the texture of citizenship." #ethics #AI #tech #democracy

On "a $1.5 million fine to settle a lawsuit that originally sought at least $150 million. If approved by a federal court, the proposed settlement submitted yesterday would require a trust in Musk’s name to pay a $1.5 million civil penalty to the government." #ethics #law #tech

"In February, The New York Times reported that Google, Reddit, Discord, and Meta had received hundreds of administrative subpoenas during the previous six months": arstechnica.com/tech-policy/20… #ethics #law #gov #internet #socialmedia #business #tech #privacy #security

"The demand for the man’s location #data was included in a request DHS issued to Google called a customs summons, which is supposed to be used to investigate issues related to importing goods and collecting customs duties." #ethics #law #gov #privacy

Why Perfect AI Alignment Is Mathematically Out Of Reach spectrum.ieee.org/ai-alignment?s…

"The conventional wisdom assumes misalignment is a bug that can eventually be removed... Our results show that the problem of alignment is not simply a lack of better data, more compute, or better engineering, but a limit built into both formal systems and universal computation."

Will humans keep making meaningful choices, or will AI decide for us? In Reclaiming Human Agency in the Age of Artificial Intelligence, @scuethics's Brian Patrick Green explores how “agentic” tech is reshaping what it means to be human. Read more: go.scu.edu/4tr4qSc