AI Advertising Company Hacked

At least some of this is coming to light:

Doublespeed, a startup backed by Andreessen Horowitz (a16z) that uses a phone farm to manage at least hundreds of AI-generated social media accounts and promote products has been hacked. The hack reveals what products the AI-generated accounts are promoting, often without the required disclosure that these are advertisements, and allowed the hacker to take control of more than 1,000 smartphones that power the company.

The hacker, who asked for anonymity because he feared retaliation from the company, said he reported the vulnerability to Doublespeed on October 31. At the time of writing, the hacker said he still has access to the company’s backend, including the phone farm itself.

Slashdot thread.

Posted on December 19, 2025 at 7:02 AM8 Comments

Comments

KC December 19, 2025 11:33 AM

I’m not on TikTok, but TikTok says the flagged accounts are now labelled as being AI-generated.

TikTok’s Community Guidelines are super interesting.

404 Media links to the section on ‘Edited Media and AI-Generated Content (AIGC)’

https://www.tiktok.com/community-guidelines/en/integrity-authenticity#3

Really basically, and I hope I have this right, creators need to label AI-gen content showing people. They also prohibit content that is misleading about matters of public importance or harmful to individuals. The details therein matter a lot (toggle More information) and I’d love to see their review process.

lurker December 19, 2025 1:04 PM

So, if the Social Media industry is corrupt, and
the Advertising industry is corrupt,
advertising on social media must invite more convoluted forms of corruption. There’s ditty about fleas on the back of other fleas …

Yet another case of technical solutions not working for a social problem.

Clive Robinson December 19, 2025 3:48 PM

@ lurker, Bruce, ALL,

Is the web becoming of less interest than politics?

It’s actually a relevant question here as the Internet income is predicated on “eyes on adverts” actually becoming sales.

If people that control personal budgets loose interest in the Internet, then sales will actually be hit disproportionately fast in a less than zero wage rise and falling employment opportunity economy.

Normally I’d pop this on the Friday Squid, but as I’ve said it’s relevant here.

Faith in the internet is fading among young Brits

Ofcom survey finds 18-34s increasingly see life online as bad for society and their mental health

The UK’s communications regulator found that in June 2025, just a third of those aged 18-34 agreed the internet is good for society, down from 42 percent a year earlier. While this fell for older age groups as well, it did so less sharply (34 percent versus 38 percent in 2024), meaning those aged 55 and above are proportionately more positive than younger people about the internet’s impact on civilization.

https://www.theregister.com/2025/12/19/internet_bad_for_society/

Thus it’s not just “falling eyeball count” but “disenchanted eyeball increase” as well.

Thus this Venture Capital Funded startup is not just running out of moat / runway it’s potential clients are not going to be seeing returns expected / promised.

Thus the question that has to be asked,

“Is this a good thing or a bad thing?”

The answer to which very much depends on your point of view.

For the social element of society I would say it’s good news. As the Social Networking Mega-Corps will get hit and their ability to purchase influence will decrease probably faster than their ad based income.

Will it reduce “enshitification” initially no it will get worse as those dependent on ad income double down. But in the longer term the web will move to new business models or in effect go bankrupt at the server end.

Unfortunately that will also help authoritarian politicians and their guard labour. Because one thing that will happen is major services will become “Pay to Play” and very very few are actually worth signing up to “Government backed Surveillance” for. Which will further reduce the “eye balls on ads”.

It will also effect politics, not just the dirty money flows endemic in many so called democracies but the ability to reach eye balls at minimum cost and convenience.

But the one thing that almost certainly will be true is based on the old observation of,

“You Can’t go back”

Especially true when it comes to “harms” of any kind, they’ve happened and the scars will go to the grave with many. All that is gained is caution and lessons to teach.

This fake influencers VC backed startup is not just breaking terms and conditions of Companies… In some jurisdictions it’s committing fraud.

For those in the UK look at the three types of fraud outlined in Sub section 2 of Section 1 of the fraud act,

https://www.legislation.gov.uk/ukpga/2006/35/section/1

It could be fairly said a case could be made under all three types of fraud…

Jeff S December 20, 2025 11:37 AM

Why should I take you seriously anymore, since you’re a signatory to the Resonant Computing Manifesto? LLMs are a plague on society, and you’re not helping.

Clive Robinson December 20, 2025 11:57 PM

@ Jeff S, ALL,

You mention the very very recent,

“Resonant Computing Manifesto”

Without providing a link/reference to it.

Like you I happen to think it’s “Surveillance friendly” as it’s way to amenable to “interpretation” thus open to the surveillance equivalent of “Greenwashing”[1].

The question thus arises,

“Who is deceiving who?”

And as with many things “new” there is often an element of “hope to do good” and consequent “optimism” that is in practice “self deceiving”.

Often this is due to a misunderstanding of the intent and motivations of what is in effect the opponent, that has often hidden motivations they are not going to change.

A century ago Upton Sinclair ran into the issue and made his now oft quoted comment of,

“It is difficult to get a man to understand something when his salary depends upon his not understanding it.”

It shows that a hidden intent will manifest in protective or prohibitive behaviours due to “self interest”.

But whilst Upton uses a single example of “salary” the same rule applies to all other forms of “self interest” especially in those with a sense of “self entitlement”. What you see as a harm against many, they see as their “unquestionable right”. Which is why I’ve often made the statement of it being a significant problem of,

“Individual Rights -v- Social Responsibilities”

But to assess this, you have to go through the proposal not just on it’s merits but why it’s got significant issues. So point out the potential and probable failings with the proposal.

Neither you nor I need “break ground” on this as others already have.

One of the gentler examples is,

https://notebook.wesleyac.com/resonant-computing/

The “Resonant Computing Manifesto” is not very good

A friend informed me recently of The Resonant Computing Manifesto — the latest from the “malleable systems”​/​“digital gardening”​/​“knowledge systems” folks, who appear to every few years need a fancy new term to describe the utopia that they’re going to build any day now. In this most recent iteration, “AI” will lead us to this utopia, but only if we are careful to follow the principles they are laying out.

In this post, I’m going to go through the entire manifesto, and point out the things that I find suspect.

I hope every one reads it and notes what they agree or disagree with and why.

My own viewpoint is that the “self entitled” will just ignore it all and pay somebody else or groups of people to become “fronting puppets” / “lobbyists” to buy time etc, whilst they carry on regardless.

Personally I have a long standing total disagreement with this sort of statement,

“With the emergence of artificial intelligence, we stand at a crossroads. This technology holds genuine promise.”

Firstly the Current AI LLM and ML Systems are not in any way “intelligent” and never can be. As I’ve explained before LLMs are nothing more than “complex filters” that can not reason, only match and add noise. Whilst the building process can work from actual rules added by humans, it can find statistical rules via the ML. However statistics of this form can not determine if it is cause or correlation, and worse the correlations found can be wrong due to the order the data is analysed in (which actually makes it a really poor statistical tool).

But consider five points,

1, Tools are agnostic to use.
2, Social problems do not have technological solutions.
3, The use is decided by a Directing Mind.
4, All actions have a defence.
5, Acts of agents are deniable.

So,

A knife is designed to cut both effectively and efficiently, so it cuts food and throats equally as well…

Cutting a throat however does not solve the underlying social problem, it just removes an easily replaceable puppet…

The choice of cutting a throat is that of the directing mind, not the knife that has no agency to act independently…

To say the act of cutting a throat is Good or Bad, is importantly entirely subjective and based on an individuals point of view of the circumstances…

One thing we’ve learned about corporations is that actions are always carried out at arms length by another entity or agent of the Directing Mind. This shields the Directing Mind because it allows for plausible deniability / scapegoating / chucking under the bus. Worse the Directing Mind does not have to be singular, the use of committee dilutes or nullifies any direct responsibility so destroys “beyond reasonable doubt”…

I could go on, but I’ve made these points long in the past, and they haf nullify this manifesto before it was even brought into the public view.

[1] Greenwashing is a term used to name the core behaviours of environmentally polluting product and service suppliers that use incredibly deceptive practices to appear as what they or their products or services are not.

A dictionary definition is,

“Greenwashing (noun) : The dissemination of misleading information that conceals abuse of the environment in order to present a positive public image.”

Clive Robinson December 21, 2025 8:50 AM

@ Bruce, ALL,

We all know time is relative, but to what is important.

The story is a little confusing, but it raises a serious question about how we,

1, Build infrastructure.
2, Build security on it.

Put simply for very many reasons computers usually have three types of “time” running on them,

1, CPU cycles elapsed.
2, RTC time hardware.
3, UTC “Wall Clock” time.

Of the three it’s maintaining the “Wall Clock” time that is usually most important.

In fact in a high dependency network system such as the modern Internet and Mobile Comms have become it’s actually critical that “wall clock” time is all synced together. Clock drift is for many things “unacceptable” which we already see with Satellites and data communications (and is going to make life less than fun when we move out off of Earth and “relativity really bites).

Which is why the “Network Time Protocol”(NTP) has become of major importance. Even though due to previous reliability it’s effectively become an “invisible critical infrastructure element” for all but a few, and various Governments around the world sponsor. Which means in effect it has produced a fairly massive set of “Master-Slave” relationships that have to be reliable on an unreliable network infrastructure…

So what happens when things “go wrong”?

To which the answer is “it depends”.

Hence this strange sounding story,

https://www.theregister.com/2025/12/21/nist_ntp_outage_warning/y6yt

NIST tried to pull the pin on NTP servers after blackout caused atomic clock drift :

A rare case of deliberately trying to induce an outage.

A staffer at the USA’s National Institute of Standards and Technology (NIST) tried to disable backup generators powering some of its Network Time Protocol infrastructure, after a power outage around Boulder, Colorado, led to errors.

The thing is it can be shown that there is no way to make the time protocol infrastructure “reliable”…

Leave a comment

Blog moderation policy

Login

Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via https://michelf.ca/projects/php-markdown/extra/

Sidebar photo of Bruce Schneier by Joe MacInnis.