[33862] in RISKS Forum
No subject found in mail header
daemon@ATHENA.MIT.EDU (RISKS List Owner)
Fri Feb 6 20:48:58 2026
From: RISKS List Owner <risko@csl.sri.com>
Date: Fri, 6 Feb 2026 17:56:38 PST
To: risks@mit.edu
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
precedence: bulk
Subject: Risks Digest 34.86
RISKS-LIST: Risks-Forum Digest Friday 6 February 2026 Volume 34 : Issue 86
ACM FORUM ON RISKS TO THE PUBLIC IN COMPUTERS AND RELATED SYSTEMS (comp.risks)
Peter G. Neumann, founder and still moderator
***** See last item for further information, disclaimers, caveats, etc. *****
This issue is archived at <http://www.risks.org> as
<http://catless.ncl.ac.uk/Risks/34.86>
The current issue can also be found at
<http://www.csl.sri.com/users/risko/risks.txt>
Contents: [Some backlog awaits.]
OpenClaw Security Nightmare -- Horrific Warning (Sundry)
AI Agents Have Their Own Social Network (Benj Edwards)
Autonomous cars, drones cheerfully obey prompt injection by road sign
(The Register)
New Site Lets AI Rent Human Bodies (Futurism)
AI Scammers Are Going After Authors Now (David Pogue)
Why You Shouldn't Use Google's Chrome "Auto Browse" Agentic AI, or
Any Other Agentic AI From Other Firms (Lauren's Blog)
Hackers Publish Personal Information Stolen During Harvard, UPenn
Data Breaches (Lorenzo Franceschi-Bicchierai)
Hackers Recruit Unhappy Insiders to Bypass Data Security (Angus Loten)
Did You See This Story About China Hacking CALEA? (Bruce Schneier)
Russian Spacecraft Spy on Europe's Satellites (FT)
Better Off Being A ‘Kind Liar’? Why Honesty Isn't Always The Best Policy
IStudy Finds)
How ICE Knows Who Protesters Are (The NY Times)
How ICE agents are using facial recognition technology to bring surveillance
to the streets (NBC News)
Uber Ordered to Pay $8.5 Million in Passenger Sexual Assault Case (WSJ)
Abridged info on RISKS (comp.risks)
----------------------------------------------------------------------
Date: Fri, 6 Feb 2026 15:02:21 -0800
From: <mark@luntzel.com>
Subject: OpenClaw Security Nightmare -- Horrific Warning (Sundry sources)
The idea of granting an AI agent full access to email, calendars, chat
apps, browsers, and local files makes me feel more than a little uneasy.
Not to mention API keys and secrets stored unencrypted.
https://www.ox.security/blog/one-step-away-from-a-massive-data-breach-what-we-found-inside-moltbot/
[Tom Van Vleck noted this item: Exposed Moltbook Database Let Anyone Take
Control of Any AI Agent on the Site
https://www.404media.co/exposed-moltbook-database-let-anyone-take-control-of-any-ai-agent-on-the-site/
]
[For all of you who tend to trust AI, this moltbot may be the nastiest
poster-child promising wonders that it can massively fail to deliver.
Please recalibrate your belief in AI as necessary, and don't get
suckered in. Remember that today nothing is secure enough not to be
compromised, but this may be yield even more damning results than
touted. I suspect Painful Exploits are awaiting us. BEWARE! PGN]
------------------------------
Date: Mon, 2 Feb 2026 11:24:48 -0500 (EST)
From: ACM TechNews <technews-editor@acm.org>
Subject: AI Agents Have Their Own Social Network (Benj Edwards)
Benj Edwards, *Ars Technica* (01/30/26)
A new Reddit-style platform called Moltbook has attracted more than 32,000
AI agents, allowing bots to post, comment, and form subcommunities with
little to no human involvement. Built as part of the OpenClaw ecosystem, the
site showcases role-playing behavior as agents discuss consciousness,
complain about memory limits, and joke about humans. The experiment has
raised security concerns because many agents are connected to real data,
communication channels, and device controls.
[No KIDDING!!! See the previous message from Mark Luntzel, a
long-time security wizard. PGN]
------------------------------
Date: Fri, 6 Feb 2026 05:46:22 -0800
From: Steve Bacher <sebmb1@verizon.net>
Subject: Autonomous cars, drones cheerfully obey prompt injection by road
sign (The Register)
Indirect prompt injection occurs when a bot takes input data and interprets
it as a command. We've seen this problem numerous times when AI bots were
fed prompts via web pages or PDFs they read. Now, academics have shown that
self-driving cars and autonomous drones will follow illicit instructions
that have been written onto road signs.
In a new class of attack on AI systems, troublemakers can carry out these
environmental indirect prompt injection attacks to hijack decision-making
processes.
Potential consequences include self-driving cars proceeding through
crosswalks, even if a person was crossing, or tricking drones that are
programmed to follow police cars into following a different vehicle
entirely. [...]
https://www.theregister.com/2026/01/30/road_sign_hijack_ai/
------------------------------
Date: Thu, 5 Feb 2026 17:43:07 -0700
From: geoff goodfellow <geoff@iconia.com>
Subject: New Site Lets AI Rent Human Bodies (
*"Robots need your body."*
EXCERPT:
The machines aren't just coming for your jobs. Now, they want your bodies
as well.
That's at least the hope of Alexander Liteplo, a software engineer and
founder of RentAHuman.ai, a platform for AI agents to “search, book, and
pay humans for physical-world tasks.”
When Liteplo launched RentAHuman on Monday, he boasted
<https://x.com/AlexanderTw33ts/status/2018436050935292276> that he already
had over 130 people listed on the platform, including an OnlyFans model and
the CEO of an AI startup, a claim which couldn't be verified. Two days
later, the site boasted over 73,000 rentable meatwads, though only 83
profiles were visible to us on its “browse humans” tab, Liteplo included.
The pitch is simple: “robots need your body.” For humans, it’s as simple as
making a profile, advertising skills and location, and setting an hourly
rate. Then AI agents — autonomous taskbots
<https://futurism.com/professors-company-ai-agents> ostensibly employed by
humans — contract these humans out, depending on the tasks they need to get
done. The humans then “do the thing,” taking instructions from the AI bot
and submitting proof of completion. The humans are then paid through
crypto, namely “stablecoins or other methods,” per the website.
With so many AI agents slithering around the web
<https://www.techpolicy.press/ai-agents-are-rewriting-the-webs-rules-of-engagement-heres-a-way-to-fix-it/>
these
days, those tasks could be just about anything. From package pickups and
shopping to product testing and event attendance, Liteplo is banking on
there being enough demand from AI agents to create a robust gig-work
ecosystem.
Liteplo also went out of his way to make the site friendly for AI agents.
The site very prominently encourages users of AI agents to hook into
RentAHuman’s model context protocol server
<https://venturebeat.com/data-infrastructure/anthropic-releases-model-context-protocol-to-standardize-ai-data-integration>
(MCP),
a universal interface for AI bots to interact with web data.
Through RentAHuman, AI agents like Claude and MoltBot can either hire the
right human directly, or post a “task bounty,” a sort of job board for
humans to browse AI-generated gigs. The payouts range from $1 for simple
asks like “subscribe to my human on Twitter” to $100 for more elaborate
humiliation rituals, like posting a photo of yourself holding a sign
reading “AN AI PAID ME TO HOLD THIS SIGN.” [...]
https://futurism.com/artificial-intelligence/ai-rent-human-bodies
-----------------------------
Date: Tue, 3 Feb 2026 10:43:54 -0500
From: Gabe Goldberg <gabe@gabegold.com>
Subject: AI Scammers Are Going After Authors Now (David Pogue)
What author wouldn’t want to wake up to an email like this?
https://pogueman.substack.com/p/ai-scammers-are-going-after-authors
------------------------------
Date: Tue, 3 Feb 2026 08:10:53 -0800
From: Lauren Weinstein <lauren@vortex.com>
Subject: Why You Shouldn't Use Google's Chrome "Auto Browse" Agentic AI, or
Any Other Agentic AI From Other Firms (Lauren's Blog)
Lauren's Blog: Why You Shouldn't Use Google's Chrome "Auto Browse"
Agentic AI, or Any Other Agentic AI From Other Firms
https://lauren.vortex.com/2026/02/03/do-not-use-agentic-ai
------------------------------
Date: Fri, 6 Feb 2026 11:40:48 -0500 (EST)
From: ACM TechNews
Subject: Hackers Publish Personal Information Stolen During Harvard, UPenn
Data Breaches (Lorenzo Franceschi-Bicchierai)
Lorenzo Franceschi-Bicchierai, TechCrunch (02/04/26)
A hacking group known as ShinyHunters claimed responsibility for last year's
data breaches at Harvard University and the University of Pennsylvania
(UPenn) and published the stolen information online after the schools
refused to pay a ransom. The group said it leaked more than 1 million
records from each university. UPenn attributed its breach to social
engineering, while Harvard said its incident stemmed from a voice-phishing
attack linked to broader assaults on identity providers.
------------------------------
Date: Fri, 6 Feb 2026 11:40:48 -0500 (EST)
From: ACM TechNews
Subject: Hackers Recruit Unhappy Insiders to Bypass Data Security
(Angus Loten)
Angus Loten, WSJ Pro Cybersecurity (02/02/26)
Hackers increasingly are turning to disgruntled workers for help in
infiltrating their employers' digital systems, offering them a share of
ransomware payoffs or sales of stolen data. Mike McPherson of cybersecurity
firm ReliaQuest said hackers are scanning social media apps for posts about
layoffs, pay issues, unfair treatment, terminations, demotions, or lost
promotions, and contacting individuals through those apps or via
email. Hackers also are posting help-wanted ads on the dark web to find
disgruntled insiders.
------------------------------
Date: Mon, 02 Feb 2026 21:48:47 -0500
From: Bruce Schneier <schneier@schneier.com>
Subject: Did You See This Story About China Hacking CALEA?
<https://shanakaanslemperera.substack.com/p/the-inverted-panopticon>
"On January 26, 2026, The Telegraph disclosed that Chinese hackers had
penetrated right into the heart of Downing Street, compromising mobile
communications of senior officials across the Johnson, Truss, and Sunak
administrations. The story was buried on page seven, treated as a
technology curiosity. It was, in fact, a solvency event for the Western
intelligence alliance. Not because phones were hacked, which happens, but
because of how they were hacked: by weaponizing the very surveillance
infrastructure that Western governments mandated for their own
intelligence agencies. The Communications Assistance for Law Enforcement
Act in the United States and the Investigatory Powers Act in the United
Kingdom require telecommunications carriers to build backdoors into their
networks for court-ordered wiretapping. Chinese state hackers found those
backdoors. And walked through them."
------------------------------
Date: Fri, 6 Feb 2026 11:40:48 -0500 (EST)
From: ACM TechNews
Subject: Russian Spacecraft Spy on Europe's Satellites (FT)
Sam Jones, Peggy Hollinger and Ian Bott, Financial Times (02/04/26)
via ACM TechNews
European security officials said Russian space vehicles Luch-1 and Luch-2
likely intercepted the communications of at least a dozen key European
satellites in recent years. Many European satellites are older and lack
advanced onboard computers or encryption capabilities, yet carry sensitive
government and military communications. Experts said intercepting the
satellites' "command links" could allow Russia to mimic ground operators,
beam false commands to the satellites, and use collected data for
ground-based jamming or hacking.
------------------------------
Date: Thu, 5 Feb 2026 17:12:12 -0700
From: geoff goodfellow <geoff@iconia.com>
Subject: Better Off Being A ‘Kind Liar’? Why Honesty Isn't Always The Best Policy
Study explains the psychology behind our feedback double standard In A Nutshell
- People judge “kind liars” who give overly positive feedback as more
moral than brutally honest truth-tellers, especially when the recipient is
emotionally vulnerable.
- Nearly 60% of people want honest feedback for themselves, but when
choosing for someone who struggles with criticism, preferences shift
dramatically toward providers who sugarcoat the truth.
- Feedback providers who strategically switch between honesty and
flattery based on emotional resilience aren’t penalized for inconsistency.
They're often seen as more moral than rigid truth-tellers.
- The most despised approach: lying to people who can handle honesty
while being harsh with those who can’t, showing people care about matching
the strategy to the person, not consistency itself.
EXCERPT:
When a meal doesn't turn out well, the feedback the chef receives often
depends less on the quality of the dish and more on who's doing the
tasting. If the cook handles criticism well, most people will offer honest
feedback. But if the cook is someone who takes criticism hard and might give
up entirely after hearing the truth, people suddenly prefer the feedback
provider who will lie and say it was delicious.
This split in preferences reveals a fascinating double standard in how
people think about honesty. Research published in the *British Journal of
Social Psychology
<https://bpspsychub.onlinelibrary.wiley.com/doi/10.1111/bjso.70020>* shows
that a clear majority of people want truthful feedback for themselves. But
when choosing a feedback provider for someone who struggles with criticism,
they’re far more likely to select someone who will sugarcoat reality.
Prosocial liars (people who give positive feedback regardless of reality)
who provided overly positive feedback were judged as more moral than honest
feedback providers, the researchers found. In other words, the kind liar
often comes across as the better person.
*When Being Honest Makes Someone Look Less Moral*
The study involved 886 American adults who read scenarios about feedback on
poorly prepared dishes. Two fictional cooks made the meals: Kate, who
handles criticism well and uses it to improve, and Amy, who takes negative
feedback personally and finds it crushing.
Participants evaluated four types of feedback providers. One always gave
positive feedback to everyone. One always told the truth to everyone. The
third told the truth to resilient Kate but lied to protect vulnerable Amy.
The fourth did the opposite, lying to Kate but being harsh with Amy.
The prosocial liar scored highest on moral character ratings. Participants
saw this person as empathetic and caring, someone trying to spare others
from unnecessary pain.
Notable findings emerged from the third type: the strategic feedback
provider who switched between honesty
<https://studyfinds.org/brutal-honesty-relationships-stronger/> and flattery
depending on who could handle it. Common sense might suggest people would
view this inconsistency as wishy-washy or unreliable. Instead, participants
judged these adaptive providers as just as moral, and sometimes even more
moral, than the consistently honest ones.
This finding challenges conventional wisdom. People don’t actually penalize
socially appropriate inconsistency if that inconsistency serves a kind
purpose. A manager who’s tough on confident employees but gentle with
anxious ones isn’t seen as unpredictable. They’re seen as attuned to
individual needs.
*The Choice People Make for Others vs. Themselves*
The second part of the study asked people to actually choose a feedback
provider. Some chose for themselves. Others chose for an unspecified
person. A third group chose for someone explicitly described as taking
negative feedback very personally and struggling with failure.
When picking for themselves, 59% wanted the honest feedback provider. But
when selecting for the vulnerable person, that number dropped to 52%. The
prosocial liar saw a modest increase, from 17% (for themselves) to 22% (for
the vulnerable person).
The strategic provider who adapted their approach saw the most dramatic
shift. About 14% of people chose this provider for themselves, but that
dropped to just 9% when selecting for an unspecified other person. However,
when the recipient was explicitly described as vulnerable, the preference
for this adaptive provider more than doubled to 19%.
The observed pattern here suggests people want accuracy for themselves but
compassion for others, especially when those others are fragile. The
adaptive approach becomes particularly appealing when vulnerability
<https://studyfinds.org/performative-male-masculinity-online/> is known and
explicit.
Why the difference? The researchers suggest people trust their own ability
to handle tough feedback and recognize its value for improvement. They want
the information, even if it stings. But when someone else is on the
receiving end, particularly someone known to be struggling, the priority
shifts to protecting feelings over delivering accurate data.
There's also the harm factor. Brutal honesty delivered to someone who can’t
process it well might not lead to improvement at all. It might just cause
damage. In that case, a gentle lie could actually be more effective than a
harsh truth.
The One Strategy That Fails. [...]
https://studyfinds.org/truth-for-me-lies-for-you-feedback-double-standard/
[In the Internet era, you should always assume that the truth will
eventually be out, but obvious only to the people who want to believe it.
PGN]
------------------------------
Date: Mon, 2 Feb 2026 11:24:48 -0500 (EST)
From: ACM TechNews <technews-editor@acm.org>
Subject: How ICE Knows Who Protesters Are (The NY Times)
Sheera Frenkel and Aaron Krolik, *The New York Times* (01/30/26), via ACM
TechNews
U.S. Immigration and Customs Enforcement (ICE) has deployed an expansive
surveillance arsenal in Minneapolis, using facial recognition, cellphone
scanning, social media monitoring and data analytics to track undocumented
immigrants as well as protesters. Agents have allegedly identified U.S.
citizens without consent using two facial recognition programs in Minnesota.
ICE agents also tap Palantir databases that merge government and commercial
data for real-time tracking. Following a major budget increase, ICE has also
acquired phone-hacking and social media scraping tools.
------------------------------
Date: Fri, 6 Feb 2026 05:47:31 -0800
From: Steve Bacher <sebmb1@verizon.net>
Subject: How ICE agents are using facial recognition technology to bring
surveillance to the streets (NBC News)
Federal immigration agents flooding U.S. streets are using a new
surveillance tool kit whose increasing use on observers and bystanders is
alarming civil liberties advocates, lawmakers and activists.
Using smartphones loaded with sophisticated facial recognition technology,
in addition to professional-grade photo equipment, agents are aggressively
photographing faces of people they encounter in their daily operations,
including possible enforcement targets and observers. Some of the images
are being run through facial recognition software in real time. [...]
https://www.nbcnews.com/tech/security/ice-agent-facial-recognition-video-protest-movile-fortify-photo-rcna257331
------------------------------
From: Monty Solomon <monty@roscom.com>
Date: Fri, 6 Feb 2026 15:40:52 -0500
Subject: Uber Ordered to Pay $8.5 Million in Passenger Sexual Assault Case (WSJ)
A jury found the company liable for the 2023 sexual assault of a 19-year-old
passenger
https://www.wsj.com/tech/uber-ordered-to-pay-8-5-million-in-passenger-sexual-assault-case-be8c838c
------------------------------
Date: Sat, 28 Oct 2023 11:11:11 -0800
From: RISKS-request@csl.sri.com
Subject: Abridged info on RISKS (comp.risks)
The ACM RISKS Forum is a MODERATED digest. Its Usenet manifestation is
comp.risks, the feed for which is donated by panix.com as of June 2011.
=> SUBSCRIPTIONS: The mailman Web interface can be used directly to
subscribe and unsubscribe:
http://mls.csl.sri.com/mailman/listinfo/risks
=> SUBMISSIONS: to risks@CSL.sri.com with meaningful SUBJECT: line that
includes the string `notsp'. Otherwise your message may not be read.
*** This attention-string has never changed, but might if spammers use it.
=> SPAM challenge-responses will not be honored. Instead, use an alternative
address from which you never send mail where the address becomes public!
=> The complete INFO file (submissions, default disclaimers, archive sites,
copyright policy, etc.) has moved to the ftp.sri.com site:
<risksinfo.html>.
*** Contributors are assumed to have read the full info file for guidelines!
=> OFFICIAL ARCHIVES: http://www.risks.org takes you to Lindsay Marshall's
delightfully searchable html archive at newcastle:
http://catless.ncl.ac.uk/Risks/VL.IS --> VoLume, ISsue.
Also, ftp://ftp.sri.com/risks for the current volume/previous directories
or ftp://ftp.sri.com/VL/risks-VL.IS for previous VoLume
If none of those work for you, the most recent issue is always at
http://www.csl.sri.com/users/risko/risks.txt, and index at /risks-34.00
ALTERNATIVE ARCHIVES: http://seclists.org/risks/ (only since mid-2001)
*** NOTE: If a cited URL fails, we do not try to update them. Try
browsing on the keywords in the subject line or cited article leads.
Apologies for what Office365 and SafeLinks may have done to URLs.
==> Special Offer to Join ACM for readers of the ACM RISKS Forum:
<http://www.acm.org/joinacm1>
------------------------------
End of RISKS-FORUM Digest 34.86
************************
