The RISKS Digest
Volume 32 Issue 36

Sunday, 8th November 2020

Forum on Risks to the Public in Computers and Related Systems

ACM Committee on Computers and Public Policy, Peter G. Neumann, moderator

Please try the URL privacy information feature enabled by clicking the flashlight icon above. This will reveal two icons after each link the body of the digest. The shield takes you to a breakdown of Terms of Service for the site - however only a small number of sites are covered at the moment. The flashlight take you to an analysis of the various trackers etc. that the linked site delivers. Please let the website maintainer know if you find this useful or not. As a RISKS reader, you will probably not be surprised by what is revealed…


Where are our self-driving cars?
Who's watching the legacy software systems?
James Paul
Whale Sculpture Stops Train From Plunge in the Netherlands
UK app failed to notify exposed citizens
The Guardian
Schools Adopt Face Recognition in the Name of Fighting Covid
CoVID and security awareness training
Rob Slade
When algorithmic fairness fixes fail: The case for keeping humans in the loop
Facial recognition used to identify Lafayette Square protester accused of assault
Driver prosecuted when relying on Tesla autopilot and collides with stationery object on motorway
Stephen Mason)
Microsoft Productivity Score and personalized experiences—here's what's new to Microsoft 365 in October
Microsoft 365 Blog via Gabe Gpldberg
Feds Seize $1 Billion in Stolen Silk Road Bitcoins
Clicked on a Malicious Mail, Fired, Charged with Fraud
Amos Shapir
What It's Like to Stress-Test Berlin's Brand New, Much Maligned Airport
Atlas Obscura
Company forced to change name that could be used to hack websites
The Guardian
Australia constructing giant 300-megawatt battery
Can robots help to save the ailing F&B industry?
Richard Stein
Responsible Military Use of Artificial Intelligence: Can the European Union Lead the Way in Developing Best Practice
SIPRI via Diego Latella
Re: Censorship or Sensibility?
John Levine and Sam Steingold
Re: Defective Panels in Solar Arrays
Henry Baker
Re: Using AI to control a camera at a sports event—oops!
Erling Kristiansen
Re: UK national police computer down for 10 hours after engineer pulled the plug
Attila the Hun
Re: Elon Musk's SpaceX says it will make its own laws on Mars
Amos Shapir
Remember—Remembrance, Thanksgiving, Armistice Day
Rob Slade
Info on RISKS (comp.risks)

Where are our self-driving cars? (

Richard Stein <>
Thu, 5 Nov 2020 11:23:39 +0800

"As long as self-driving features require the driver to be ready to take
control, the driver will remain liable for any accidents. Car manufacturers
are only liable if there's a fault in their vehicle. But what happens if an
autonomous passenger car causes an accident? Is the manufacturer liable
because it designed the system that's at fault?

"Some states are trying to address the question. Florida passed a law saying
that the person who initiates a trip in an autonomous vehicle is considered
the operator, and while the law doesn't explicitly establish liability, it
is laying a foundation for how liability may be addressed.  But the process
is piecemeal, and so far existing laws haven't faced serious challenges in

Given the consumer appeal for app-hailing transport, why shouldn't Florida's
taxpayers underwrite DV accident liability?

Carl Hiassen, the noted author of numerous satirical adventures about the
Sunshine State, anticipates eventual DV deployment:

  "The Florida in my novels is not as seedy as the real Florida. It's hard
  to stay ahead of the curve. Every time I write a scene that I think is the
  sickest thing I have ever dreamed up, it is surpassed by something that
  happens in real life." --Carl Hiassen


Who's watching the legacy software systems?

James Paul <>
Fri, 6 Nov 2020 22:29:01 -0500
The European journal /New Scientist /has an article in its 7 Nov 2020 issue
("Code Red").

It describes the pandemic driving new and unexpected loads on systems like
those managing state unemployment benefits distribution.  Managers find
themselves with legacy COBOL code and no staff able to repair errors as they
crop up. It's a familiar tale for RISKS readers, with fresh examples.

Whale Sculpture Stops Train From Plunge in the Netherlands (NYTimes)

Gabe Goldberg <>
Wed, 4 Nov 2020 12:27:02 -0500

The immediate risk? Gravity...

UK app failed to notify exposed citizens

Don Gilman <>
Fri, 6 Nov 2020 15:50:37 -0600

Schools Adopt Face Recognition in the Name of Fighting Covid (WiReD)

Gabe Goldberg <>
Wed, 4 Nov 2020 14:38:53 -0500
A WIRED investigation finds dozens of districts have purchased thermal
cameras to monitor fevers that can also identify students and staff.

CoVID and security awareness training

Rob Slade <>
Thu, 5 Nov 2020 12:35:47 -0800
One of the security lessons from CoVID-19 that didn't make it into the book
is in regard to security awareness training.

I have long been an advocate of education in general, and security awareness
training in particular.  A number of people say that security awareness
training doesn't work.  I maintain that security awareness training doesn't
*always* work, but, in most cases, those who say it doesn't work really
haven't actually tried doing it.  It's hard to win an argument like that,
since there is so little evidence one way or another.

Well, now we have evidence.

Many jurisdictions have tried various ways of controlling the pandemic.
Some have lockdowns, some have enforced lockdowns, some have nothing, some
have red zones, some have various types of orders to do or not do certain
things.  In BC we do have various health orders.  But our chief medical
health officer, Dr. Bonnie Henry, has constantly stressed the utility of
education and support over shutdown orders or mask mandates.  "The
Dr. Bonnie Show (co-starring Adrian Dix)" used to be daily, although now
it's down to twice a week.  But every briefing our journalists in BC
(proving how resistant *they* are to learning) ask a variation on the same
question: "why aren't you more vicious in cracking down on/forbidding house
parties, beach parties, people who don't wear masks, school re-openings, and
other people who do things I don't like?"  And every briefing Dr. Bonnie
says the same thing: education is more effective and is working better than
bullying.  And she's actually doing the "awareness training" in regard to
the pandemic.

In BC, we are doing OK.  We aren't clear, by any means.  We are undergoing a
worrying surge in cases right now.  (Two of my grandchildren are part of
that surge, so, for me, it's very worrying.  Although they do seem to be
getting better.)  We are doing better than Ontario and Quebec, even when
adjusted by population.  We are doing better than Alberta, which has a
smaller population.  We are doing better than the US.  (Well, who isn't?)
We aren't doing as well as New Zealand, but who is?  We are doing better
than Sweden, which seems to hurt my case, except that Martin seems to be
indicating that Sweden's "education" seems to consist of "we are the
government and know best, so do what we say."

As Dr. Bonnie says (pretty much every briefing, in response to the daily
"viciousness" question), most of those in BC are trying to do the right
thing.  At the mall yesterday I saw a woman who had been sitting at a table
wipe down, with a disinfecting wipe, the whole thing as she was leaving.
She told me she always does.  She also *wasn't* wearing a mask, although she
*was* trying to keep distanced from people.  So, no, awareness isn't
perfect.  But it does do something, and it does seem to be keeping our
numbers lower than in other places.  Without imposing complete lockdowns.

So, try some awareness training.

And wash your hands.

When algorithmic fairness fixes fail: The case for keeping humans in the loop (

Richard Stein <>
Mon, 9 Nov 2020 10:53:58 +0800

This article discusses healthcare service allocation versus cost
optimization using machine-driven decisions to deny/approve delivery.  The
expenses incurred to deliver a product or service to a customer confronts
all businesses.

"Pfohl agrees: 'I would argue that if you're in a setting where making a
prediction doesn't allow you to help people better, then you have to
question the use of machine learning, period. You have to step back and
solve a different problem or not solve the problem at all.'"

Human labor comprises a significant business expense. Delegate decision
approval/denial to a machine programmed to optimize service allocation given
their cost and, voila, expenditures shrink. Profits can rise autonomously
without humans-in-the-loop.

In healthcare service allocation, using patient profile characteristics
(body mass index, blood chemistry, gender, ethnicity, etc.) as key
discriminators for treatment approval/denial can elevate unsatisfactory
treatment outcome risk frequency. Unsatisfactory patient (or customer)
outcomes, especially for large cohorts experiencing profile-driven
discrimination, promotes lawsuits.

Corporate bottom lines increasingly reveal that human v. machine competition
favors algorithmic efficiency—machines—to enable profit pursuit. Weak
regulations and selective enforcement encourages this corporate strategy.

Businesses can generate profit by training an effective workforce: a
resilient talent pool of fewer employees possessing interdisciplinary
skills.  Mechanized operations, with human oversight in the loop, can
function profitably when professionally, responsibly, and ethically managed.

Risk: Unsupervised machine-based profit capture

Facial recognition used to identify Lafayette Square protester accused of assault (WashPost)

Monty Solomon <>
Tue, 3 Nov 2020 14:32:06 -0500
Justin Jouvenal and Spencer S. Hsu, 2 Nov 2020

The protester might never have been identified, but an officer found an
image of the man on Twitter and investigators fed it into a facial
recognition system, court documents state. They found a match and made an

The court documents are believed to be the first public acknowledgment that
authorities used the controversial technology in connection with the widely
criticized sweep of largely peaceful protesters ahead of a photo op by
President Trump. The case is one of a growing number nationwide in which
authorities have turned to facial recognition software to help identify
protesters accused of violence.

The case also provides the first detailed look at a powerful new regional
facial recognition system that officials said has been used more than 12,000
times since 2019 and contains a database of 1.4 million people but operates
almost entirely outside the public view. Fourteen local and federal agencies
have access.

Public defenders, defense attorneys and facial recognition experts said they
were unaware of the existence of the National Capital Region Facial
Recognition Investigative Leads System (NCRFRILS). Several said the
Lafayette Square case was the first time they had seen its use disclosed to
a defendant despite thousands of searches in bank robberies, human
trafficking and gang cases. [...]

Driver prosecuted when relying on Tesla autopilot and collides with stationery object on motorway

Stephen Mason <>
Fri, 06 Nov 2020 17:01:19 +0000
I thought your readers might be interested in the 4 articles we have
published on the Post Office Horizon scandal this year:

  Peter Bernard Ladkin, Bev Littlewood, Harold Thimbleby and Martyn Thomas
  CBE, The Law Commission presumption concerning the dependability of
  computer evidence, 17 Digital Evidence and Electronic Signature Law Review
  (2020) 1 14.

  Peter Bernard Ladkin, 'Robustness of software, 17 Digital Evidence and
  Electronic Signature Law Review (2020) 15 24

  Paul Marshall, The harm that judges do—misunderstanding computer
  evidence: Mr Castleton's story, 17 Digital Evidence and Electronic
  Signature Law Review (2020) 25 48

  James Christie, The Post Office Horizon IT scandal and the presumption of
  the dependability of computer evidence, 17 Digital Evidence and Electronic
  Signature Law Review (2020) 49 70.

The prosecution:

  Name of case: PEN 17 16 DIP, Regionalgericht Emmental-Oberaargau,
  Strafabteilung (Regional Court Emmental-Oberaargau, Criminal Division),
  30 May 2018

  Switzerland; criminal law; traffic violation; Autobahn; Tesla motor
  vehicle Traffic-Aware Cruise Control, and Autosteer mode engaged;
  collision; driver failed to control vehicle; Convention on Road Traffic,
  Vienna; [?]  value of report by Tesla Motors Switzerland GmbH URL:

Microsoft Productivity Score and personalized experiences—here's what's new to Microsoft 365 in October (Microsoft 365 Blog)

Gabe Goldberg <>
Sun, 8 Nov 2020 16:46:28 -0500
Power your digital transformation

This month, we are announcing two new offerings to help power your
transformation digital: Microsoft Productivity Score and Microsoft Cloud for

Measure and improve how your organization leverages Microsoft 365 to get
work done: It's essential that people have the tools they need to do their
best work. But tools alone are not enough, you also need to help everyone in
your organization build the habits that harness the true power of those
tools. Up until now, it's been difficult for leaders to get insight into
these habits, and to understand how to help people make the most of the
technology they invest in. Productivity Score can help by giving you
visibility into how your organization works, insights to identify where you
can make improvements, and actions you can take to update skills and systems
so that everyone can do their best work. To get started, open your
Productivity Score in the Microsoft 365 Admin Center.

The risks? Marketing blather and new/improved snoopware?

Feds Seize $1 Billion in Stolen Silk Road Bitcoins (WiReD)

Gabe Goldberg <>
Sun, 8 Nov 2020 01:31:15 -0500
A hacker identified only as Individual X had been sitting on a
cryptocurrency gold mine for seven years before the IRS came knocking.

Clicked on a Malicious Mail, Fired, Charged with Fraud ()

Amos Shapir <>
Sun, 8 Nov 2020 11:06:17 +0200
A worker of a small company in Israel received mail, supposedly from her
boss, telling her to "click on this link to avoid your mail account being
canceled".  She did, and followed instructions to enter her username and
password.  Nothing happened, so she forgot about it.

A few months later, the company's bank account received several fraudulent
requests for payment to a firm in Malaysia, backed by invoices supposedly
approved by this worker.  She was fired, and charged with fraud.  Neither
the police nor the prosecution cared about her claims of phishing.

She was acquitted, and won a false firing claim, when it was proven that
other workers in the same company—including the manager—had received
similar phishing mails, and that management had neglected to warn workers
about it, change passwords, etc.

Source (Hebrew):

What It's Like to Stress-Test Berlin's Brand New, Much Maligned Airport (Atlas Obscura)

Gabe Goldberg <>
Sun, 8 Nov 2020 00:13:16 -0500
If this doesn't sound very German, you're right, and you've probably never
lived in Berlin—which has a reputation in Germany for chaos and
incompetence that BER only reinforced. Over the last eight years, the
terminal's troubles have provided Germany's capital and its leaders with
steady servings of humiliation, and its taxpayers with a new target for
their already renowned black humor:

  What do BER and Mars have in common? It's possible people will first land
  on both of them 30 years from now.

 Wouldn't it be cheaper to tear down Berlin and rebuild it next to a
 functioning airport?

Company forced to change name that could be used to hack websites (The Guardian)

Neil Padgen <>
Sat, 7 Nov 2020 10:38:15 +0000
A software engineer was able to register a UK company name which would
expose XSS vulnerabilities if listed on improperly-secured sites.

Companies House forced the company to change its name.  It is now legally

The risk? You can never trust any input from an external source, even if
that source should be highly trustworthy.

Australia constructing giant 300-megawatt battery (

Richard Stein <>
Sat, 7 Nov 2020 14:01:41 +0800

"Victoria officials say consumers should expect to see a return of $2 for
every dollar invested in the project. The state will pay Neoem $84 million
for the power grid."

Lithium battery storage that buffer electrical distribution systems directly
dependent on renewable sources are cost-effective for small-scale
deployment. Wholesale replacement of carbon-source energy generation with a
renewable source using lithium battery storage is apparently too expensive:
manufacturing and life cycle maintenance expense are prohibitive.

Discussion of peak-power supplement to natural-gas generator facilities
using lithium battery storage can be found here:
(retrieved on 07NOV2020).

The reference conveniently estimates the funding necessary to entirely
replace California's carbon energy generation capacity with renewables that
incorporates lithium battery storage infrastructure for excess power. The
expense is staggering.

A nationwide embrace of renewable (non-carbon) energy sources, per Japan's
recent pledge to adopt ammonia
retrieved on 07NOV2020) is very bold. Ammonia leaks are hazardous.
shows that Japan will have to swap-out ~500 Terawatt-hours of carbon-energy
generation capacity over the next 40 years to achieve a ~90% green energy

There are clean energy storage mechanisms (carbon footprint-wise) that are
apparently less expensive to operate: potential energy (train payloads at
high altitude rolling down hill), molten sodium, or liquefied air. Each has
advantages and disadvantages. Scientific and engineering judgment can
identify the most effective solutions.  Persuading politicians to embrace
these facts will challenge generations.

Can robots help to save the ailing F&B industry?

Richard Stein <>
Tue, 3 Nov 2020 15:56:06 +0800

"They can mix a cocktail as fast as a human bartender can, and make coffee
that tastes almost as good as a master's brew, but can they address the
declining footfall in F&B outlets?"

What food and beverage establishment owner would turn down this innovation?

The boozebot minimizes waste, doesn't earn a salary, no healthcare insurance
to pay, ready-to-mix-and-pour out-of-the-box, etc. A little WD-40, and an
occasional stepper-motor tuneup suffices to keep the liquor and soft drinks

To rope in F&B foot traffic, an enterprising roboticist might equip their
boozebot product with an ELIZA-like voice-chatterbot to engage customer
banter. Loyal imbibers can opt-out of automatic digital psychological
profile construction, voice-print capture, etc., the perfect nightcap after
a night on the town.

"Sometimes a man'll tell his bartender things he'll never tell his doctor."
--Dr. Phil Boyce in "Star Trek, The Cage"


Responsible Military Use of Artificial Intelligence: Can the European Union Lead the Way in Developing Best Practice (SIPRI)

"Diego.Latella" <>
Wed, 04 Nov 2020 18:06:09 +0100
You might be interested in the following SIPRI report

Responsible Military Use of Artificial Intelligence: Can the European Union
Lead the Way in Developing Best Practice? [1]

Vincent Boulanin, Netta Goussac , Laura Bruun and Luke Richards SIPRI [2]
November 2020

Accessible also from the USPID ( [3] ) in the page on
Computers: National Security, War, and Civil Rights
( )


Re: Censorship or Sensibility? (Steingold, RISKS-32.35)

"John Levine" <>
2 Nov 2020 18:14:32 -0500
> CDA230 created a 3rd option for communications providers: in addition to
> "wire providers" (think ATT: no control over content, no responsibility
> for it) and "information providers" (think CNN: full control over content,
> full responsibility), we now have FB/Twitter/Google who have full control
> and no responsibility.

That is so wrong I do not know where to start. For one thing, the liability
model in 230 is not new. It is the one that has applied to newsstands and
bookstores forever. Nobody expects the owner of a bookstore to know what is
in every book, and if it turns out some of the books have illegal content
(historically meaning pornography) the store owner isn't liable.

The newsstand rule was applied to Internet providers by a Federal court
under Cubby vs. Compuserve in 1991. But then, a NY state court misread that
decision in Stratton Oakmont vs. Prodigy in 1995, which led Prodigy's owner
IBM to push for the protection in the CDA which allowed them to moderate
their user discussions without subjecting themselves to impossible legal
standards. (Ironically, the allegedly defamatory posts saying that Stratton
Oakmont was a fraud were true, and its managers later went to jail.)

This same protection from 230 applies to every discussion list, including
this one, and every web site that has user comments. It is literally true
that without 230, the Internet would be nothing like it is now with only a
tiny fraction of the discussions, debates, and user comments we expect.

For much more on this see Mike Godwin's recent article here:

> How about applying CDA230 only to _small_ players?  If you have more than
> 10% of all US users, you cannot censor content.

Ah, so Facebook has to deliver every phish, every bit of anti-science
nonsense, and every bit of spam? That is surely just what their users need
and want.

  [Sam Steingold responded: and right now FB stops "anti-science nonsense"
  that does not suit their political agenda, but keeps spreading
  "anti-science nonsense" that supports it.
  This would certainly be a dramatic improvement compared with the current
  state of affairs where all users see is the news approved by the Agitprop
  Committee (Google/FB/Twitter).  We live in a world of media monopoly which
  will have disastrous consequences - "democracy dies in darkness".  This
  reminds me of USSR ca. 1970-ies & 80-ies: you can say whatever you want to
  your friends in the kitchen, but all the media is centrally controlled.
  You cannot spread an unapproved message - either in the USSR (jail time if
  you try) or in today's USA (monopolies stop the message and cancel culture
  destroys your life).
  Even if I agreed with the monopolies' political agenda, I would have still
  opposed the current situation because it destroys free exchange of ideas.]

  [Combine the responses as interstitiation makes more sense rather than
  having to be repetitious in order to understand a reply to a replies.

Re: Defective Panels in Solar Arrays (Ladkin, RISKS-32.35)

Henry Baker <>
Mon, 02 Nov 2020 16:46:31 -0800
Re: Mulilo Sonnedix Prieska solar farm panel failures

I don't know about this particular failure, but some of the older solar
panel systems utilize *direct current* at relatively high voltage—just
the sort of current one might utilize for *arc lighting* or *arc welding*!
Indeed, check out some of the YouTube videos about solar panel system
failures with DC arc-ing.

Many of the newer solar systems are based upon *alternating current* --
typically of the same sort that you're already using—e.g., 120VAC,
240VAC, etc.  While AC systems require multiple *inverters*, these inverters
pay for themselves through better optimization of the energy output from
each panel separately, and through added resilience due to redundancy (think
RAIP—Redundant Array of Independent Panels; parallel Xmas tree lights
rather than the old serial Xmas tree lights).

The good news is that Moore's Law makes the inverters and computer controls
cheaper with time; eventually, I expect to see inverters integrated into
every individual solar panel.

So the age-old Tesla/Edison battle of the currents continues to rage.

BTW, the efficiency of most solar panels is higher when they are kept cooler
-- e.g., by enabling air to circulate underneath.  I'm surprised that solar
panels don't come with integrated solar powered *fans*—I think that the
trivial energy to power the fan is more than repaid by the higher output
from the cooler panel. I would guess that the panel lifetime is also
improved through cooler operation.

Re: Using AI to control a camera at a sports event—oops! (RISKS 32.35)

Erling Kristiansen <>
Wed, 4 Nov 2020 11:55:44 +0100
Despite the name, AI is not really intelligent at all, and, in particular,
it is missing the context that would prevent a human camera operator from
making such a mistake.

Re: UK national police computer down for 10 hours after engineer pulled the plug (RISKS-32.35)

Attila the Hun <>
Wed, 4 Nov 2020 11:21:51 +0000
Dick Mills says that he can easily imagine an even bigger outcry if other
certain systems were found to be impossible to switch off by the actions of
a single person.

Without sight of those "other certain" systems I cannot be definite, but
experience strongly suggests that *any* outcry would be either limited to
specialist media or a small puff of wind in a demitasse.

But it reminded me of a joke that went around the Atlas computer lab in
Manchester University (UK) in the late 1960s:

  British computer-maker ICL had designed and built the world's most
  powerful computer [I did say it was a joke], and the Queen was invited to
  inaugurate it.  She was offered the opportunity to ask it a question and
  after a moment's contemplation she asked, "Is there a God?" [a question of
  obvious interest to a lady with the title 'Fidei Defensor'].

  The console lights flashed, the tape decks whirled, Friden Flexowriters
  chattered, Creed tape punches spewed out ribbons of perforated paper,
  sprocket-fed stationary poured from the back of Analex printers and clouds
  of Freon gas gushed from vents in the many steel-grey cabinets.  After
  some time it was clear that Her Majesty was becoming restless, but just as
  the MD of ICL leaned forward to explain, all the console lights went out
  and the room fell silent.  Clearly the decisive moment was at hand [the
  young Douglas Adams would have drawn inspiration from the drama].

  Suddenly the overhead lights went out as well, plunging the room and the
  royal party into Stygian blackness.  Then, with a crack like a whip, from
  one of the still off-gassing cabinets there leapt a bolt of lightning,
  straight at the main circuit-breaker, welding it shut.

  The console teletype clattered into action, typing out the words:

Re: Elon Musk's SpaceX says it will make its own laws on Mars (RISKS-32.35)

Amos Shapir <>
Sun, 8 Nov 2020 09:37:49 +0200
This reminds me of Robert Heinlein's sci-fi novel "Stranger in a Strange
Land", where the first person born on Mars is determined to be, by some
legal quirk, the legal owner and king of Mars.

Remember—Remembrance, Thanksgiving, Armistice Day

Rob Slade <>
Tue, 3 Nov 2020 10:51:58 -0800
To my friends in the Unexplored Southern Area:

I hope your election goes well today.

However, I'm a bit more concerned that, regardless of what happens with the
election, you guys are on track to hit 10 million CoVID cases, probably by
Remembrance Day.  (Which seems somehow rather hideously appropriate.)

Take warning by us: our Thanksgiving Day is earlier than yours.  About a
week later, we had the beginnings of a surge that is still going on.
Including two of my grandchildren.  And your Thanksgiving is a bigger deal
than ours is.  This is *NOT* the year to go home for the holidays.  Figure
out some creative way to celebrate *without* getting different people from
different households into close proximity.

Anyway, given that Remembrance/Armistice Day is coming up, herewith a little
early this year:

Terry Kelly's "A Pittance of Time"

Please report problems with the web pages to the maintainer