, 40 tweets, 8 min read Read on Twitter
@CommonsCMS heard from the Electoral Commission, The Institute for Strategic Dialogue and two technology companies on issues relating to Electoral integrity and also donations including via PayPal

BXP has high risk systems

Here is a link to the session.

parliamentlive.tv/Event/Index/2b…
2/ EC said that PayPal is used & can be used safely and properly providing political parties ensure their obligations are met.

Some parties use PayPal safely with systems & training so traceable Know Your Donor (KYD) processes are established and applied.

Brexit Party not so.
3/. The EC did not believe naivety was the reason in BXP case given they had politicians & personnel that had worked within other political parties.

The EC also ran oversight on PV, OFOC & CHUK but did not find the same high risk systems.
PayPal configured by them appropriately
4/. Those configured systems better enabled them to check the source of donations and aggregations of smaller (under £500) donations.

Recommendations have been made to the BXP that means they have to be able to demonstrate they can trace the source and aggregation of donations
5/. If they cannot do so then the money has to be returned within 30 days of donation.

Failure to comply means that the money is forfeit to the EC and then the Treasury.

EC says they will check “in a few weeks” that the work has been done.

Why so long?
6/. The party Registered on 5/2/2019.

That means there have been several months of potential donations (& according to the BXP these amounted to £ millions).

That is 4 months.
7/. @IanCLucas asked if it was shown that the funds cannot be provided to have come from permissible sources would that then render the EU19 results void.

The EC says “NO”. The only sanction is forfeiture and fines of £20k per offence. Clearly NOT a deterrent.

Unsatisfactory.
8/. In addition @IanCLucas asked if there was a pre operation certification system whereby political,parties had to receive certification from the EC that they had robust systems, staffing and training in place before being allowed to participate in elections.

EC. No.
9/. This would require legislation.

Yet recommendations made in EC reports to Gov dating back to 2014 and 2018 have yet to be implemented by Government.

Posner (EC) also says that cherry picking bits of the recommendations is undesirable as they work as a coherent platform.
10/. Had they been implemented many of the issues that arose particularly in the EU19s would have been significantly mitigated although nothing could compensate for the very late notice given to the EC by the Government of participation in the EU19s
11/. Here is a link to the FOI made to the Electoral Commission in relation to funding of The Brexit Party, CHUK, PV and OFOC.

It also contains the Electoral Commission response and recommendations.

electoralcommission.org.uk/__data/assets/…
12/
@JoStevensLabour asked the EC what work they undertook to establish the confidence the public had in the Electoral systems.

They do survey the public and the findings were that there was lower satisfaction with both sufficiency of rules & effectiveness of enforcement
13/. At present Anti- Money laundering legislation has not been extended to political parties to ensure a similar level,of proactive compliance.

Thought should be given to changing the leglislation in that regard.
14/. Turning to Social Media.

Whilst adverts now bear an imprint of a person or organisation The imprints do not necessarily disclose who is really paying for it.

The information about targeting not sufficiently granular. New regs should improve this.
16/. The Institute for Strategic Dialogue gave evidence.

During EU 19s they mapped

Social media analytics
Smaller technology platforms ..chat rooms...WhatsApp.
Network mapping
Also looking at state level media. Eg. Russia. Hungary. Poland
17/. Trends in disinformation has become much more complex. Hostile state actor still deployed but also more complex transnational non state actors, extreme and populous actors.

Now harder to distinguish legitimate and illegitimate information deployment.
18/. There has been a move away from fake news distributed through bot nets.
Now legitimate sources being distributed out of context.

Immigration. Family values. Anti feminist. Anti LGBT. Climate denial.

Actors multiple. Often aligned with US Special interest groups.
19/. Targets long term Information networks as an operation to shift cultural norms

So Notre Dame quickly used to leverage anti Muslim sentiment.

Remigration dialogue. “The great replacement” (eg Christchurch massacre) breaking into the mainstream ..eg German AfD Party.
20/. ISD working with Univeristies (eg MIT) to trace political language to see the point at which a narrative builds momentum. So moving from fringe groups to wider population. Track the proliferation of ideas and compare to polling data. Hyperactivity. Track against polling
21/. There is a need to shift from content moderation eg just removing illegal material towards looking at algorithms and the architecture of platforms.
22/. Look at Problematic behaviours inc changes in Bots/Sock puppets.
The system being gamed by shifting approaches.

It is possible to buy networks of existing accounts for almost no money make them more difficult to spot. Then used to power content proliferation.
23/. It is very plain that a voluntary code of conduct simply is not working.

ISD said they provided Facebook with weekly reports of bad actor activity in the 4 weeks up to the EU elections.

They only received one response with a question but no answers or report back.
24/. They said they know that some PARTS of a malign network AVAAZ identified to FB were acted on but not all.

They asked FB for information on malign operation. False or distortive information and networks.

No satisfactory feedback.
25/. They saw explicit bot accounts on Twitter.

42% of the top 10 accounts centring around each Party for the EU19 looked at gave strong indications of botlike activity.

They check not just frequency of tweeting but also how authentic the account seems (Name/Identifiable)
26/.

The top 10 accounts deploying for the Brexit Party were most probably bot accounts.
27/. @IanCLucas says there was a false statement made about him on FB. Raised with FB. It refused to take it down.

The same issue in relation to the falsified Pelosi video. FB did not remove. YouTube did.

FB community standards say that defamation is impermissible.
28/. Tom Packalen (ex police) from UTOPIA ANALYTICS, a Finnish company delivering MACHINE LEARNING & TEXT Analytics. All languages and dialects.

Used to identify in millisecond suspect text for action eg. Child grooming, hate speech, terrorism.
29/. He says, during the Sri Lankan they offered to build a model in 2 weeks for FB using the local dialect.

Facebook were not interested in working with 3rd Parties.

Yet there were 250 deaths.
30/. Who applies the rule about what should be removed?

The test analysis company provides a tool and the company determines how it should be applied. They use what has been applied before to formulate the tool. NLP
31/. The Institute for Strategic Dialogue thinks there is potential in this approach. It is already used in some instances- 35% for hate speech. 95% re Terrorist content.

A lot depends upon the company priority.

Different issues re image and video moderation.
32/. Research suggests that people working in a company (eg FB) moderating materials are given 8 seconds to make decisions to review content...and can cause the moderator PTSD.

There is a problem for 3rd party moderation inc Govts in that we cannot see behind the wall.
33/. It is worth looking at the JUNE 2019 Grand Parliament hearing
- 14 Nations Hearing in Canada. Zuckerberg and Sandberg NA

It really is quite chilling. 20 highlights from @jason_kint

34/. Philip Gradwell from CHAIN ANALYSIS gave evidence on
Blockchain
He asserts it is transparent: BUT the transactions are pseudo anonymous - No name
Chain analysis provides the mapping of the systems
Not seeing a lot of crypto currency flowing into U.K. political parties.
34/. More elsewhere crypto currency seen elsewhere and maybe
a bit more evidence on far right groups. Likely to see more in the future.

Important to have human systems to ensure Know Your Donor can be established securely.
Sorry, that was 35/.

36/. But good at tracking donations (including micro donations) to source . The final stage of linking to a named human being or company has to be done through other means and may require court order and cross jurisdiction cooperation.

Russia?!
37/. It is known that Facebook is planning its own crypto currency.
It has a large user base & that increases the risk of bad actors. So definitely a risk. Anti money laundering systems needed.
BUT Can trace then multiple small donations to a single source/wallet.
38/. This is about the business models being applied too.

The Finnish Company, Utopia Analytics, did point out that if their text analysis resulted in 5% of posts being removed then that, in turn, given the business model used by SM platforms, would reduce advertising revenue.
39/. Further when the SM company deliberately applies algorythms to drive people towards specified sites to increase their revenue that has to be capable of being policed, not just self-policed that is not working.

Here is an older thread to explain why

threader.app/thread/1107298…
40/. An update in today about Facebook’s intended cryptocurrency.

Given Facebook’s blatant data abuses and signal failure to police their own platforms effectively I am beginning to think it should be shut down completely whilst Governments design effective management oversight,
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Fionna O'Leary #FBPE
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!