Political Milestones for AI

ChatGPT was released just nine months ago, and we are still learning how it will affect our daily lives, our careers, and even our systems of self-governance.

But when it comes to how AI may threaten our democracy, much of the public conversation lacks imagination. People talk about the danger of campaigns that attack opponents with fake images (or fake audio or video) because we already have decades of experience dealing with doctored images. We’re on the lookout for foreign governments that spread misinformation because we were traumatized by the 2016 US presidential election. And we worry that AI-generated opinions will swamp the political preferences of real people because we’ve seen political “astroturfing”—the use of fake online accounts to give the illusion of support for a policy—grow for decades.

Threats of this sort seem urgent and disturbing because they’re salient. We know what to look for, and we can easily imagine their effects.

The truth is, the future will be much more interesting. And even some of the most stupendous potential impacts of AI on politics won’t be all bad. We can draw some fairly straight lines between the current capabilities of AI tools and real-world outcomes that, by the standards of current public understanding, seem truly startling.

With this in mind, we propose six milestones that will herald a new era of democratic politics driven by AI. All feel achievable—perhaps not with today’s technology and levels of AI adoption, but very possibly in the near future.

Good benchmarks should be meaningful, representing significant outcomes that come with real-world consequences. They should be plausible; they must be realistically achievable in the foreseeable future. And they should be observable—we should be able to recognize when they’ve been achieved.

Worries about AI swaying an election will very likely fail the observability test. While the risks of election manipulation through the robotic promotion of a candidate’s or party’s interests is a legitimate threat, elections are massively complex. Just as the debate continues to rage over why and how Donald Trump won the presidency in 2016, we’re unlikely to be able to attribute a surprising electoral outcome to any particular AI intervention.

Thinking further into the future: Could an AI candidate ever be elected to office? In the world of speculative fiction, from The Twilight Zone to Black Mirror, there is growing interest in the possibility of an AI or technologically assisted, otherwise-not-traditionally-eligible candidate winning an election. In an era where deepfaked videos can misrepresent the views and actions of human candidates and human politicians can choose to be represented by AI avatars or even robots, it is certainly possible for an AI candidate to mimic the media presence of a politician. Virtual politicians have received votes in national elections, for example in Russia in 2017. But this doesn’t pass the plausibility test. The voting public and legal establishment are likely to accept more and more automation and assistance supported by AI, but the age of non-human elected officials is far off.

Let’s start with some milestones that are already on the cusp of reality. These are achievements that seem well within the technical scope of existing AI technologies and for which the groundwork has already been laid.

Milestone #1: The acceptance by a legislature or agency of a testimony or comment generated by, and submitted under the name of, an AI.

Arguably, we’ve already seen legislation drafted by AI, albeit under the direction of human users and introduced by human legislators. After some early examples of bills written by AIs were introduced in Massachusetts and the US House of Representatives, many major legislative bodies have had their “first bill written by AI,” “used ChatGPT to generate committee remarks,” or “first floor speech written by AI” events.

Many of these bills and speeches are more stunt than serious, and they have received more criticism than consideration. They are short, have trivial levels of policy substance, or were heavily edited or guided by human legislators (through highly specific prompts to large language model-based AI tools like ChatGPT).

The interesting milestone along these lines will be the acceptance of testimony on legislation, or a comment submitted to an agency, drafted entirely by AI. To be sure, a large fraction of all writing going forward will be assisted by—and will truly benefit from—AI assistive technologies. So to avoid making this milestone trivial, we have to add the second clause: “submitted under the name of the AI.”

What would make this benchmark significant is the submission under the AI’s own name; that is, the acceptance by a governing body of the AI as proffering a legitimate perspective in public debate. Regardless of the public fervor over AI, this one won’t take long. The New York Times has published a letter under the name of ChatGPT (responding to an opinion piece we wrote), and legislators are already turning to AI to write high-profile opening remarks at committee hearings.

Milestone #2: The adoption of the first novel legislative amendment to a bill written by AI.

Moving beyond testimony, there is an immediate pathway for AI-generated policies to become law: microlegislation. This involves making tweaks to existing laws or bills that are tuned to serve some particular interest. It is a natural starting point for AI because it’s tightly scoped, involving small changes guided by a clear directive associated with a well-defined purpose.

By design, microlegislation is often implemented surreptitiously. It may even be filed anonymously within a deluge of other amendments to obscure its intended beneficiary. For that reason, microlegislation can often be bad for society, and it is ripe for exploitation by generative AI that would otherwise be subject to heavy scrutiny from a polity on guard for risks posed by AI.

Milestone #3: AI-generated political messaging outscores campaign consultant recommendations in poll testing.

Some of the most important near-term implications of AI for politics will happen largely behind closed doors. Like everyone else, political campaigners and pollsters will turn to AI to help with their jobs. We’re already seeing campaigners turn to AI-generated images to manufacture social content and pollsters simulate results using AI-generated respondents.

The next step in this evolution is political messaging developed by AI. A mainstay of the campaigner’s toolbox today is the message testing survey, where a few alternate formulations of a position are written down and tested with audiences to see which will generate more attention and a more positive response. Just as an experienced political pollster can anticipate effective messaging strategies pretty well based on observations from past campaigns and their impression of the state of the public debate, so can an AI trained on reams of public discourse, campaign rhetoric, and political reporting.

With these near-term milestones firmly in sight, let’s look further to some truly revolutionary possibilities. While these concepts may have seemed absurd just a year ago, they are increasingly conceivable with either current or near-future technologies.

Milestone #4: AI creates a political party with its own platform, attracting human candidates who win elections.

While an AI is unlikely to be allowed to run for and hold office, it is plausible that one may be able to found a political party. An AI could generate a political platform calculated to attract the interest of some cross-section of the public and, acting independently or through a human intermediary (hired help, like a political consultant or legal firm), could register formally as a political party. It could collect signatures to win a place on ballots and attract human candidates to run for office under its banner.

A big step in this direction has already been taken, via the campaign of the Danish Synthetic Party in 2022. An artist collective in Denmark created an AI chatbot to interact with human members of its community on Discord, exploring political ideology in conversation with them and on the basis of an analysis of historical party platforms in the country. All this happened with earlier generations of general purpose AI, not current systems like ChatGPT. However, the party failed to receive enough signatures to earn a spot on the ballot, and therefore did not win parliamentary representation.

Future AI-led efforts may succeed. One could imagine a generative AI with skills at the level of or beyond today’s leading technologies could formulate a set of policy positions targeted to build support among people of a specific demographic, or even an effective consensus platform capable of attracting broad-based support. Particularly in a European-style multiparty system, we can imagine a new party with a strong news hook—an AI at its core—winning attention and votes.

Milestone #5: AI autonomously generates profit and makes political campaign contributions.

Let’s turn next to the essential capability of modern politics: fundraising. “An entity capable of directing contributions to a campaign fund” might be a realpolitik definition of a political actor, and AI is potentially capable of this.

Like a human, an AI could conceivably generate contributions to a political campaign in a variety of ways. It could take a seed investment from a human controlling the AI and invest it to yield a return. It could start a business that generates revenue. There is growing interest and experimentation in auto-hustling: AI agents that set about autonomously growing businesses or otherwise generating profit. While ChatGPT-generated businesses may not yet have taken the world by storm, this possibility is in the same spirit as the algorithmic agents powering modern high-speed trading and so-called autonomous finance capabilities that are already helping to automate business and financial decisions.

Or, like most political entrepreneurs, AI could generate political messaging to convince humans to spend their own money on a defined campaign or cause. The AI would likely need to have some humans in the loop, and register its activities to the government (in the US context, as officers of a 501(c)(4) or political action committee).

Milestone #6: AI achieves a coordinated policy outcome across multiple jurisdictions.

Lastly, we come to the most meaningful of impacts: achieving outcomes in public policy. Even if AI cannot—now or in the future—be said to have its own desires or preferences, it could be programmed by humans to have a goal, such as lowering taxes or relieving a market regulation.

An AI has many of the same tools humans use to achieve these ends. It may advocate, formulating messaging and promoting ideas through digital channels like social media posts and videos. It may lobby, directing ideas and influence to key policymakers, even writing legislation. It may spend; see milestone #5.

The “multiple jurisdictions” piece is key to this milestone. A single law passed may be reasonably attributed to myriad factors: a charismatic champion, a political movement, a change in circumstances. The influence of any one actor, such as an AI, will be more demonstrable if it is successful simultaneously in many different places. And the digital scalability of AI gives it a special advantage in achieving these kinds of coordinated outcomes.

The greatest challenge to most of these milestones is their observability: will we know it when we see it? The first campaign consultant whose ideas lose out to an AI may not be eager to report that fact. Neither will the campaign. Regarding fundraising, it’s hard enough for us to track down the human actors who are responsible for the “dark money” contributions controlling much of modern political finance; will we know if a future dominant force in fundraising for political action committees is an AI?

We’re likely to observe some of these milestones indirectly. At some point, perhaps politicians’ dollars will start migrating en masse to AI-based campaign consultancies and, eventually, we may realize that political movements sweeping across states or countries have been AI-assisted.

While the progression of technology is often unsettling, we need not fear these milestones. A new political platform that wins public support is itself a neutral proposition; it may lead to good or bad policy outcomes. Likewise, a successful policy program may or may not be beneficial to one group of constituents or another.

We think the six milestones outlined here are among the most viable and meaningful upcoming interactions between AI and democracy, but they are hardly the only scenarios to consider. The point is that our AI-driven political future will involve far more than deepfaked campaign ads and manufactured letter-writing campaigns. We should all be thinking more creatively about what comes next and be vigilant in steering our politics toward the best possible ends, no matter their means.

This essay was written with Nathan Sanders, and previously appeared in MIT Technology Review.

Posted on August 4, 2023 at 7:07 AM33 Comments


Keith Douglas August 4, 2023 9:11 AM

Idea for #0: A major program in a significant department quietly changes to a largely AI-mediated rather human mediated one. (I suspect that in a way this has already happened.)

Al Sneed August 4, 2023 10:09 AM

we were traumatized by the 2016 US presidential election.

I’m guessing that by “we” you mean the Democratic Party here. I’m tired of the absolute COPE that Russia somehow hijacked our election with $200k of Facebook ads. If that’s the case, then AI will definitely bring about chaos and destruction and the end of “our” democracy.

But I suspect this is not the case and the blob is simply seething about losing to a literal reality TV star and a general buffoon. The type of person that doesn’t mesh with the rest of the elites and “should not” be in power.

Here’s how AI can help democracy today: make a chatbot that summarizes Congress transcripts and/or corporate earnings calls. Maybe something that also scrapes the news and simply summarizes what was said. You could even ask it questions. Of course, the news will be AI-generated soon and automatically cleaned of “misinformation” or whatever, but the call transcripts will hopefully stay reliable.

Winter August 4, 2023 10:38 AM

@Al Sneed

I’m tired of the absolute COPE that Russia somehow hijacked our election with $200k of Facebook ads.

What about Bengazi and her emails and lock her up? And things like grab her by the ????

And that was just before the elections.

Chelloveck August 4, 2023 10:48 AM

I don’t usually disagree with you so strongly, Bruce. You’re pre-supposing an AI that has agency and can act without being prompted by a human. Once we get to that point, your milestones make sense. Right now, however, that’s still a long way away. I’d wager that there’s still at least one revolutionary discovery between us and general AI; I don’t think the current crop of statistical parlor tricks is going to lead to that development. You may as well be discussing milestones for exploring the solar system using anti-gravity drives for all the relevance it has to existing technology.

Al Sneed August 4, 2023 10:53 AM

What about it? Seems like your standard political shitflinging. Russian troll farms buying Facebook ads is like pissing into an ocean of piss. Same with AI-generated content.

In general, I think we went wrong when we started taking things posted on the internet too seriously. Anyone can post anything online, and it’s about time that we come to terms with this. I suppose the screens that boomers used to get information from were more trustworthy (heh). Well, at least the narrative was more coherent across the board.

Winter August 4, 2023 11:13 AM

@Al Sneed

What about it?

Calling for incarceration of your political opponent, was not really customary in the US. And all the nonsense about Bengazi and emails proved to be pure propaganda with zero follow up

Russian troll farms buying Facebook ads is like pissing into an ocean of piss.

Russia supporting a US presidential candidate was also not customary. It was clear that the GOP was very much aligned with Putin (also financially).

Ted August 4, 2023 11:26 AM

re: milestone 5

Here comes a new boutique of ChatGPT millionaires retiring as politicians 😄

JonKnowsNothing August 4, 2023 12:08 PM

They only potentially useful outcome is if AI taught politicians how to be Great Orators.

Oratory is a specialist area in speech making.

Religious persons may learn to do it as part of their training. Some learn not to do so, because it interferes with the Standard Book of Accepted Words.

Politicians and Playwrights may have a innate ability to make great speeches, and having that power of speech is mesmerizing. (1) It remains a reason why some speeches are on the embargo list.

Oratory and Iconography can be intertwined where one boosts the other. Some examples of this are also on the embargo list.

Most politicians shy away from Oratory because they prefer their constituents to be dull, bored and uninterested. Being a brown bug on a brown background has a survival benefit.

It the not too distant past, in certain places, people assembled for an 8-10 speech. It was the cinema of the day. Large crowds assembled to hear the speech. If the speech was average or low the crowds left thankful that their required attendance was over. If the speech was great a revolution might start.

The only useful thing a politician might learn from AI is

  • How Not To Bore Your Constituents When Actually Speaking for More than 5min (2)


1) A great speech was given by DiFi about the CIA Torture documents from the floor of the Senate. Too bad it was all bluff and lies.

2) I attended a speech by a VP of the USA, it lasted less than 5min. I don’t remember a single thing that was said. Speech WAI.

lurker August 4, 2023 5:26 PM


#1: AI submitting work in its own name

This presupposes. as @Chelloveck observes, an autonomous AI capable of original thought. Current devices are far from achieving this. Yet we have people taking the current machines’ output and signing their human names to it. This muddying of the waters is not going to help progress.

Sommers August 4, 2023 5:40 PM

… Dozens of text paragraphs to say what exactly ? No obvious bottom line.

Why 6 Milestones rather than 7 … or 3 , or 15 ?
Why are such arbitrary Milestones of any importance at all ?

Simple AI could at least edit this long screed down to a more readable level.

Robert Kenneth Smart August 4, 2023 7:40 PM

AI doesn’t have motivation. It should never be allowed to have the sort of motivation that leads to independent action.
Humans don’t get motivation from intelligence. Computer programs/AI can get arbitrarily intelligent without ever acquiring the sort of motivation that would make them dangerous. Human motivation is weird and complex and interesting. It arose in evolutionary conditions which no longer exist.
Giving AI motivation wouldn’t be so bad if there was the slightest chance of it being done right. It will be easy to give them enough motivation to wipe humanity out, but then, instead of being our successors they will just do something stupid and follow us to extinction.
Giving AI motivation to act independently in the human world will be a crime against humanity of the highest order. Particularly: AI must not be allowed to pretend to be human when interacting with us.

JonKnowsNothing August 4, 2023 10:10 PM

@Robert Kenneth Smart, All

re: AI motivation

AI is nothing more than a stream of digital zeros and ones. There is no motivation at all. It only alters the stream of zeros and ones according to a formula managed by humans. The formula may have a loop effect, or do while-do until aspect, but there is no motivation involved.

It is nothing more than


The concern is that the PIXELS are arranged in a way that humans interpret the layout and color arrangement on the screen as


It’s an area of study called Metaphysics.

  • Do you believe in Tinker Bell?
  • Do you believe that the Petit Prince successfully protected his rose?
  • Do you believe Pixels On A Screen because they form letters or symbols in a defined order?

Yildo August 4, 2023 10:29 PM

If a photograph taken by a monkey cannot be attributed to the monkey, text hallucinated by neural network cannot be attributed to the neural network. None of the current breakthroughs are on the path of such “AI” having any agency whatsoever

denton scratch August 7, 2023 6:13 AM

Milestone n: an AI that can review proposed legislation, in the light of domain knowledge about the field the proposal applies to, and of related fields; identifies unintended consequences, perverse incentives, conflicts with existing legislation, hidden costs, etc.; and then produces a revised draft, along with explanatory notes.

Milestone n+1: an AI that can translate legislation into normal English that ordinary people can understand.

It infuriates me that legislators require citizens to obey laws they can’t possibly hope to understand, without paying a month’s salary to a law firm. “Ignorance of the law is no defence”, they say. If that’s the case, then only lawyers should be subject to the law.

Petre Peter August 7, 2023 8:37 AM

“Meaningful” : Information conveyed through language that can be inferred by the receiver–Siri Knowledge

“Plausible” : having an appearance of truth or reason; seemingly worthy of approval or acceptance; credible; believable: a plausible excuse; a plausible plot. –Siri Knowledge

“achievable” : able to be brought about or reached successfully. –some dictionary

“observable” : in physics, an observable is a physical quantity that can be measured. Examples include position and momentum. In systems governed by classical mechanics, it is a real-valued “function” on the set of all possible system states. In quantum physics, it is an operator, or gauge, where the property of the quantum state can be determined by some sequence of operations. For example, these operations might involve submitting the system to various electromagnetic fields and eventually reading a value. –(Wikipedia) through Siri Knowledge.

One problem with legislation written by AIs is the issue of reputation which can be the determining factor when it comes to trust. Therefore, “submitted under the name of the AI” in essence becomes a debate on the reputation of the AI. Milestone #4, “AI creates a political party with its own platform, attracting human candidates who win elections” puts a human face on the AI which will be beneficial to the relation between human and machine.

Gert-Jan August 7, 2023 2:25 PM

Disclaimer: I am not a lawyer

From a legal point of view, there is no such thing as an independent machine / AI. There’s always a human (or corporation) responsible.

I don’t think we (the human race) will ever accept anything that is not submitted or vouched for by a human.

(I realize that even that can create temporary problems, when AIs prey on vulnerable humans and them do their bidding)

There is another problem with accepting autonomous AIs. Because it is trivially easy to clone an AI. Now you have two AI that slowly drift apart (assuming the AI keeps learning). Now which AI are you trusting. What proposal comes from which AI and how would you know? How do you know the AI still exists one minute after it created a proposal? An AI may implode or explode countless identities as fast as hardware allows. Such AIs will be ghosts. I don’t think we will ever accept independent machines / AI that don’t offer accountability. I think it needs humans to vouch for that.

Clive Robinson August 7, 2023 3:06 PM

@ Gert-Jan, ALL,

“I don’t think we will ever accept independent machines / AI that don’t offer accountability.”

In the US to certain extent they already have.

Ask yourself the question,

“If a self driving vehicle causes an accident where does the legal liability exist?”

JonKnowsNothing August 7, 2023 4:03 PM

@Clive, @ Gert-Jan, ALL

re: “If a self driving vehicle causes an accident where does the legal liability exist?”


I watched a documentary about food safety in the USA, although it applies to other countries too. It isn’t new news that many items in markets can carry bacteria and viruses that make people or animals sick after touching or eating them. There have been attempts at reducing food borne illnesses but much diligence is lacking.

One of the interesting claims in the film was, that according to the US Agencies that supervise the food chain here (there are more than a dozen), the outlook on diseases often found in ground meats and chicken, is that:

  • It is the responsibility of the housewife to cook it properly

That statement might cause a few jaws to drop but it seems to be the case.

Cooking a lettuce salad is problematic.



Swedish mountain lodge closes as stomach bug spreads among hikers

A popular lodge on Sweden’s highest peak has been forced to temporarily close after a stomach bug rapidly spread among hikers.

The precise nature of the illness is yet to be determined.

The station will be fully sanitised and is expected to reopen

h ttps://www.theguardian.c o m/world/2023/aug/06/swedish-mountain-lodge-closes-as-stomach-bug-spreads-among-hikers

(url fractured)

vas pup August 7, 2023 6:02 PM

AI in medicine some extracts just to trigger Your interest


“Rhema Vaithianathan, the director of the Centre for Social Data Analytics and a professor at the Auckland University of Technology in New Zealand, who focuses on tech in health and welfare, thinks it is right that people are asking AI to help make big decisions. “We should be addressing problems that clinicians find really hard,” she says.

One of the projects she is working on involves a teen mental-health service, where young people are diagnosed and treated for self-harming behaviors. There is high demand for the clinic, and so it needs to maintain a high turnover, discharging patients as
soon as possible so that more can be brought in.

Doctors face the difficult choice between keeping existing patients in care and treating new ones. “Clinicians don’t discharge people because they’re super scared of them self-harming,” says Vaithianathan. “That’s their nightmare scenario.”

Vaithianathan and her colleagues have tried to develop a machine-learning model that can predict which patients are most at risk of future self-harming behavior and which are not, using a wide range of data, including health records and demographic information, to give doctors an additional resource in their decision-­making. “I’m always looking for those cases where a clinician is struggling and would appreciate an algorithm,” she says.

!!!And the models offer a veneer of objectivity that can lead people to pass the buck on ethical decisions, trusting the machine rather than questioning its output.

This ongoing problem is a theme in David Robinson’s new book, Voices in the Code, about the democratization of AI. Robinson, a visiting scholar at the Social Science Matrix at the University of California, Berkeley, and a member of the faculty of Apple

The prejudices at the time led the committee to favor married men with jobs and families, for example.

The way Robinson tells it, the lesson we should take from Scribner’s work is that certain processes—bureaucratic, technical, and algorithmic—can make difficult questions
seem neutral and objective. They can obscure the moral aspects of a choice—and the sometimes awful consequences.

Quantification can be a moral anesthetic, and computers make that anesthetic easier than ever to administer.”

For now, the auditing of algorithms by independent parties is more wish-list item than standard practice. But, again using the example of kidney disease, Robinson shows how it can be done.

The algorithm had been designed to allocate kidneys in a way that maximized years of life saved. This favored younger, wealthier, and whiter patients, Grawe and other patients argued.

!!!Creating a program that can assess someone’s mental health is an unsolved problem—and a controversial one. As Nitschke himself notes, doctors do not agree on what it means for a person of sound mind to choose to die. “You can get a dozen different
answers from a dozen different psychiatrists,” he says. In other words, there is no common ground on which an algorithm could even be built.

!!!For Robinson, devising algorithms is a bit like legislation: “In a certain light, the question of how best to make software code that will govern people is just a special case of how best to make laws. People disagree about the merits of different ways of
making high-stakes software, just as they disagree about the merits of different ways of making laws.” And it is people—in the broadest sense—who are ultimately responsible for the laws we have.”

Clive Robinson August 7, 2023 6:28 PM

@ JonKnowsNothing,

“It is the responsibility of the housewife to cook it properly”

Food safety, is and always has been the responsibility of the person preparing it for consumption.

Except in the case of precooked foods that are eaten cold like potted meats, charcuterie and paté. Or only have to be re-heated.

One of the scary things I’ve mentioned before are the poisonous foods we commonly eat.

Take tomatoes, whilst it’s OK to eat the ripe fruit, you don’t want to be eating or cooking the green fruit except from certain varieties, and you never want to cook with or eat the vines.

Closely related is potatoes you don’t want to be cooking or eating anything green or purple. It’s a relative of the deadly nightshade which is inedible.

In fact with most “root vegtables” you don’t want to cook with or eat the above soil plants or anytging green.

A lot of TV Chefs appear unaware of this and they use the green parts of a number of root vegtables etc in salads…

Then there are beans and peas. Whilst you can eat raw peas don’t eat the plant. A lot of beans you realy want to soak for half a day at least then bring to a “rolling boil” for ten mins then slow cook for an hour. Lentils especially those that have not been polished are one of those you should “cook to death” not just to make them digestable, but also to break down the trypsin inhibitor.

Then there are fruits and similar. Strawberries leaves are not poisonous but they can upset your gut. Be careful though because not all berries that look like strawberries are strawberries. All fruit seeds such as in apples, pears and stoned fruits contain cyanide, which you can “smell” the bitter almond smell of. Then there are the odd ones like rhubarb, whilst it’s OK to eat the stems when nolonger green you should never eat the leaves or bulb.

Fun one to look up is tapioca pudding and “Death by Casava”. There are two types of casava sweet and bitter both contain cyanide and are thus poisonous so need carefull preperation (repeated,soaking washing, and beating over several days). Bitter casava has upto fifty times as much and is a heck of a lot more poisonous than sweet casava and although it can be made safe to eat it requires a lot of work, time and water. Trust me when I say the work required to make even sweet casava edible, is realy more effort than you’ld want to put in for a starch that is lets just say an aquired taste. It’s easier to buy the prepared pearls and flours, that also have the advantage of being gluten free. You might have heard of tapioca pudding and even been given it to eat when you were young half a century or more ago. Well tapioca is basically made from casava and sugar, and very high in calories.

Oh you can if you prepare and cook them properly eat casava leaves as that is where the protien is to be found. Which is why many south American casava recipies contain the leaves.

Another one to watch out for is the Jamaican “national dish” ingredient ackee fruit. Vegitarians are taking an interest in the fruit for various reasons, but if not properly ripe it can make you very unwell in a most unpleasent way (hypoglycemic shock). The onset of which is called Jamaican vomiting sickness,


A similar issue has arisen with it’s close relative often found in Chinese food the lychee.

There’s quite a few more, but those are the ones I was taught as a kid or came to know about through friends who’s parents / grand parents had taught them.

JonKnowsNothing August 7, 2023 7:08 PM


re: Food safety, is and always has been the responsibility of the person preparing it for consumption v producer’s liability

I may not have expressed it well but the documentary highlighted poor food production practices that enhance the spread of pathogens. These practices are not hygienic to begin with and increases the spread of pathogens in the production line model of high speed and fast gather mechanics used in the USA.


Per the documentary, the spread of several pathogens can be traced directly to the production end of the process and not to the homemaker cooking it.

The reality is with today’s super bugs, home cooking is not enough to prevent illness. The packaging, the food, the fixed ingredients are already contaminated before you get the package open. Once opened in a home setting, the pathways to future contamination are many and some pathogens last years once dry on a surface.

The producers of food products know perfectly well that their production lines and techniques spread huge amounts of pathogens into the food system.

  • They swap their producers liability to a home cook to clean up the disease infused products

The old canard that it’s the responsibility of the cook, stems from long ago ~1900, when food contamination was even worse. It was devised as a way to reduce liability and production costs, increase profits, with a side helping of shuttling business into the medical industry.

The newest bug variants are deadly and cannot be killed using common home kitchen methods.

Clive Robinson August 8, 2023 5:07 AM

@ JonKnowsNothing,

Re : Supply Chain Security failings.

“… the documentary highlighted poor food production practices that enhance the spread of pathogens.”

The same notion of liability applies, in that it becomes “buyer beware” no matter how many ISO9000 etc certifications there are.

Food processing is without doubt a “dirty business” “brought in from outside” and even vegtables grown in your own back garden carry significant health risks and contamination risks, thus caution & care should be the watch words.

It is a supply chain model that should be studied as it tells us much about what can and does get done even though it should not legally, morally, or ethically.

Opperating a “dirty to clean” process that also processes is difficult and time consuming and so is expensive, thus it “eats profit” or “margins” or at the end of the day “bonuses” (see baby formular scandal in US).

As a process it’s almost as old as mankind and goes back to hunter gather nomadic life styles, so many millennia ago that what we know of it is gathered from the “dirty side”. It did not change much when mankind moved to farming and settled and became industrialized.

What realy changed it was two World Wars and the development of rapid and inexpensive mass transportation in the late 1950’s. But it took time for the process to get from bulk shipping of feedstock as carcus meat with processing at the point of final sale / consumption, to the current industrialized distributed processing across many thousands of miles and many national boarders.

It used to be when I was younger you could walk into your local greengrocer or butchers and your nose and eyes would tell you much of what you needed to know. TV dinners and the like were not common and outside of baking and confectionary food prep/cook was done in the home, by most of the family in one way or another. Family meals ment much more than eating together.

So in your High St Shop you would see things like your butcher would make ground meat to order, and likewise cut meat from the carcus to order, in front of you. Thus minimising contamination risk.

These days it’s processed in an untracable way in an unknown part of the world and shipped. All part of the “Global Supply Chain”, along with thousands of tonnes of plastic and cardboard packaging destined for the refuse system. With each layer of packaging a different “ID Code” that is supposedly tracable but meaningless more than one step further down the supply chain. Thus security not ensurable.

That’s how the UK and Europe had it’s “Horse meat burger” scandle just over a decade ago.


If anything it’s got less tracable, less secure, more open to fraud and worse since.

It’s worth investigating because it tells you what to expect in other supply chains that the ICT industry is 100% dependent on and thus fails to be even close to secure.

JonKnowsNothing August 8, 2023 1:50 PM


re: The Global Food Unsafe Supply

It’s the same problem as any other attack-defense security issue, but with the add on that sometimes people have no choice but to eat tainted food.

There is an entire range of problems: short-term & long-term all intertwined because somethings that appear to be short-term end up being long-term due to how the food is stored, canned, jarred, dried, preserved etc.

  • ASF (1) in pigs has a long environmental footprint, more than decades later it can infect the pig family. While ASF does not infect humans, you cannot see the virus by looking at the package of sausage.
  • E. coli O157:H7 (2) a major cause of modern food borne illness cannot be seen, felt, tasted on food. It’s the Tears of Lys (3) in our modern food production and delivery system. A toxin-poison powerful enough to kill.

Within different aspects of the delivery system, there are preventatives that can be done to reduce the extent of contamination. It is just not good enough to say It’s Your Own Fault and shrug off the damage done while claiming It is not possible to control because it is possible to control. It does require a change in production line methods and the primary impediment isn’t science or in engineering a better production line.

Like every aspect of security, there are trade offs: pro and con. We saw a lot of trade offs in play since 2019+ and many of those haven’t changed.

  • Dealing with contaminated greens, fruits, vegetable does have a remedy.
  • Dealing with contaminated meats, primarily ground meat and chicken has a different remedy.

The problem is:

  • Who is going to pay the piper to reduce the extensive invisible diseases on food?

Right now:

  • You pay the Farmer, the Rancher, the Market, the Doctor, the Hospital for getting sick from something they processed into the food.

YMMV ATM folks might want to consider swapping out some foods for others. Steak is less contaminated than Ground Beef and Green Salad may not be worth dying for.


1) African swine fever virus (ASFV) is a large, double-stranded DNA virus in the Asfarviridae family.[1] It is the causative agent of African swine fever (ASF). The virus causes a hemorrhagic fever with high mortality rates in domestic pigs; some isolates can cause death of animals as quickly as a week after infection. It persistently infects its natural hosts, warthogs, bushpigs, and soft ticks of the genus Ornithodoros, which likely act as a vector, with no disease signs.[2] It does not cause disease in humans.

2) Escherichia coli O157:H7 is a serotype of the bacterial species Escherichia coli and is one of the Shiga-like toxin–producing types of E. coli. It is a cause of disease, typically foodborne illness, through consumption of contaminated and raw food, including raw milk and undercooked ground beef.[1][2] Infection with this type of pathogenic bacteria may lead to hemorrhagic diarrhea, and to kidney failure; these have been reported to cause the deaths of children younger than five years of age, of elderly patients, and of patients whose immune systems are otherwise compromised.

Transmission is via the fecal–oral route, and most illness has been through distribution of contaminated raw leaf green vegetables, undercooked meat and raw milk.

3) Tears of Lys – a fictitious poison in Game of Thrones literature, that is tasteless, colorless and odorless.

Anonymous August 8, 2023 4:52 PM

#5 seems insane: who creates a bot to make money and send it to a political campaign?

lurker August 8, 2023 5:57 PM

@Clive Robinson, JonKnowsNothing

Some clarification might be needed re vegetables.

Green tomato is commonly used in pickles and chutneys, and in Latin American cuisine.

You were presumably referring to potato tops, but purple potatoes are highly sought after in some places for their ability to grow in alpine climates, and for the anti-oxidants and other good things in the purple skin (like “sweet potato”, which isn’t a potato). Potatoes which have gone green through exposure to sunlight taste bitter but are not poisonous to male humans. The green bits should be cut away because even when cooked they are implicated in fetal spina bifida for humans in the first trimester of pregnancy.

The green tops of radish, turnip, swede and mangel are excellent salad greens, but many people find them more appetising when fast cooked in stir fry. The latest food fad of microgreens includes sprouting the seeds of these varieties.

Pea tendrils, the first two leaves, and the flower, are a nutritious addition to salads, or as a garnish to other cooked dishes. Green beans also are perfectly edible raw. I assume you meant the dried haricot or lima type, which need the long slow boil to get water in to hydrolyse the protein and make it digestible. But even this cooking does not eliminate the other problem of beans: gas production during digestion. The Chinese 1600 years ago found a way to fix that: make the beans into tofu, unfortunately another time and labour intensive process.

Sweet cassava that has been freshly dug ( up to 3 hours ago) can be easily peeled, briefly rinsed in running water, cut into chunks and boiled like potatoes. I have used this method under instruction from locals in Central Africa. They told me the trick is to get the skin and cambium layer off quickly after harvest, without cutting the starchy flesh. Making flour is another industrial process.

Of course I am talking about fresh vegetables, from our own garden, or neighbours, or perhaps a peasant market. Not the “fresh vegetable” section of a supermarket. As @Clive said

It used to be when I was younger you could walk into your local greengrocer …

So we now have a problem that the supply chain for modern urban food distribution is a security risk, in its abilty to deliver quantity and quality. Can AI fix it for us by answering @JKN’s question

Who is going to pay the piper to reduce the extensive invisible diseases on food?

Clive Robinson August 8, 2023 11:30 PM

@ lurker, ALL,

Re : Veg rules.

The rules I’ve given should “keep you safe” untill you’ve learnt the other rules.

What I said about green tomato’s is a “general rule” follow it and you should not have problems.

However as I noted some spieces are safe to eat green but not all,

That said many plant poisons are alkaloid based and a lot will break down if cooked “enough”. Others if mixed with an acid such as vinigar or lemon juice will break down. Which is why,

“Green tomato is commonly used in pickles and chutneys,”

Cooked for an hour or more when making pickles/chutneys or other “savory preserves” that have vinigar in makes many safe to eat.

“You were presumably referring to potato tops, but purple potatoes are highly sought after”

Anything that grows out of a potato should be treated with suspicion, as “green potato poison” (solain) neurotoxin is found in the non tuber parts normally, including the “eyes” that grow out that start as white growths with dark purple tips. Any tubers exposed to light will also produce the neurotoxin and not only does cooking not break it down, it effects all humans,


Some of the chemicals in “greens” like oxalic acid don’t poison you directly, however they can stop you absorbing nutrients like essential minerals like calcium. In vegans this can be a significant problem. Because although spinich is high in calcium it’s also high in oxalic acid and has effectively no lipids so little or none of the calcium gets absorbed. The issue gets worse the more long chain carbohydrates –fiber– there are. So spinich and certain beans will not be at all nutritious. But oxalic acid is a known associate of “stones” in the kidneys etc in susceptible people…

So again “general rule” says avoid.

Then there are the brassicas cabages brussels, mustard plants etc, they contain a quite potent defence mechanism called the “mustard oil bomb” which can kill herbivores. Boiling cabage etc –but not other cooking methods like steaming– breaks part of the mechanism down so the isothiocyanate or thiocyanate is not produced. The problem is if you have a thiroid condition some brassicas will cause you real issues if you eat them when they have not been boild…

So again “general rule” says avoid.

Beans don’t need to be dried to be a serious problem, eating a couple of raw kidney beans can kill you as can other beans and telling the beans appart can require an expert eye just as it does with mushrooms. So the beans have to be vigorously boiled which they already have been with tinned beans… There were something like fifty near fatalities in the UK because people were not using tinned beans but “soaking and slow cooking” which does not break the poison down…

The problem with “pea tips” is like “fern tips” and some “hedgrow herbs” including nettles is when they are OK or not OK to eat and why. It’s way to complex to talk about and books have been written…

Then there are fruits… Quite a few start of poisonous then when ripe are edible, but the “pits” should not be chewed. This is the way the plant uses the dietary system of an animal to disperse it’s seeds. In fact some seeds will not grow unless they have had a hydrochloric acid soaking –stomach acid– to remove the jelly like substance around them that stops them germinating. Tomatoes are an example of this, but you can also get rid of their jelly by puting them in water and waiting for a mold to form on them after a few days, then wash then and dry them.

JonKnowsNothing August 9, 2023 12:30 AM

@ lurker, @Clive, ALL

re: Food Safe Problems

There are a number of facets to consider when evaluating the problems of safe food.

  • The food itself
  • The food as delivered

@Clive has given many good suggestions about the The Food Itself.

  • The Food Itself

There are not that many food stuffs humans eat but it can be highly regional in selection. There are types of beans that are not common in the USA but are the preferred type in Jamaica, Cuba, Haiti, and other countries. Preparation and methods of eating it vary too. We learn how to prepare our normal diet from family members. Preparing foods that are not within our normal eating pattern needs careful preparation and understanding of how to do it.

  • Getting the wrong mushroom is the last mushroom
  • The Food as Delivered

This is the area of my current interest. The Food Itself is fine to eat for most people. It’s the way the food in harvested, prepared, packaged, stored and handled that may introduce pathogens that are not inherent in the Food Itself.

  • Eating a leafy green salad, lettuce harvested same day, however, grown downwind from a cattle yard might lead to an unpleasant stay in the hospital and worse case might be the last leafy green salad you eat.

I don’t know the answer about “green tomatoes” however in the USA we have pickled green tomatoes, fried green tomatoes and you can store green tomatoes in a cool spot and they will ripen over the winter and they will turn red when ready to eat.

We also have an item that looks like a green tomato but is a tomatillo; used in making green salsa, green chili verde sauce, and is commonly added to recipes from Mexico or many Latin dishes.

There are many varieties of “banana”. Most folks in the USA eat the standard peel-it-eat it kind. A plantain which appears the same shape as the peel-it-eat-it type is heavier, less sweet and cooked, and used the same way as a potato or it can have a syrup added and eaten as dessert.

So some foods, look alike but are totally different and require different preparation.

Regardless of the type of food, reducing pathogens before eating is a prime goal of the preparation. As @Clive noted, previously it was the Cook’s Job to prepare the meal safely. Today, it’s everyone’s job to be sure what you eat is not only Prepared Properly but Presented Safely.

  • You cannot tell by visual inspection if a food item has bacteria, viruses or other pathogens.

Winter August 9, 2023 4:27 AM

Re: Food, red tomatoes and green potatoes

Contrary to some all natural folklore, almost all plants are either inedible or poisonous.[1] Selected fruits and shoots can be eaten, but most of our plant based foods have either been bred to become edible or require specific preparations to become edible.

Without fire, humans would starve. This is well illustrated by any raw food fanatics. Even with electric food processors they tend to look on the brink of starvation.[2]

Tomatoes and potatoes are members of a plant family that is almost all highly poisonous. Tomatoes and potatoes are the only parts of the respective plants that are not poisonous.

[1] Almost all vertebrates and most other animals are perfectly edible. There is the occasional fish or amphibian that is highly poisonous, but they are rare.
Fungi and mushrooms are a mixed bag in this respect.

[2] ‘https://www.sciencedirect.com/science/article/pii/S266726852300013X

Clive Robinson August 9, 2023 6:32 AM

@ JonKnowsNothing, ALL,

Re : What it tells us about ICTsec.

“Preparing foods that are not within our normal eating pattern needs careful preparation and understanding of how to do it.”

These days it gets fancy terms such as “Food Safety”, “Bio Safety” and similar along with “Supply Chain Security”, under “Food Security” as a critical element in “National Security”.

This is because of issues that came up during and following two world wars. We hear about U-Boots and Convoys but from the way it gets taught most assume it was a WWII issue[1]. The fact is that research during WWI for bio weapons[2] and poisons that attacked livestock started as an extension to attacking the food supply.

The current hostilities at the east of Europe are again about attacking food supplies. However this time due to the supply chains that were built up it is now on a global basis.

Likewise some are starting to see the conflict as the first real example of conflict based cyber-warfare.

In both cases it is the fragility of supply chains that are actually being attacked as a means to an end.

Thus whilst the last century we saw the application of science against civilian populations with the development of NBC weapons and delivery systems, a hundred years later we are seeing the application of information weapons against the global population.

All though many in ICTsec don’t think of it we are now in many cases the “front line defence troops”.

Our job is to protect entire national populations of millions of civilians. Who mostly are unknowingly critically dependent on infrastructure and logistics that even most in ICTsec don’t realise just how fragile they are.

Protecting much of the industrialised world from just a few hundred people around the globe, whos identities and locations are unknown and kill by press of an enter key.

To be able to do so effectively we have to understand how the supply chains function and fail. The best historical examples are without doubt those involving the basic necessities that have previously been targets in hostilities.

[1] Some are aware that there is an island off of the Scottish Coast “Gruinard Island” that for many years was called “Anthrax Island” where biological weapons were tested and found to be fairly inefective. What is less well known is that “green potato poison” was considered to be a bio-weapon against “herbivores not humans” such as cattle used in a nations food production. As with most biological and chemical weapons whilst the agent works the delivery mechanisms fail to live upto requirments.

[2] Whilst many are aware of the use of chemical weapons during WWI less is known about the use of bio-weapons (virus infected rats),


Likewise the deliberate targeting of civilian populations,


Post Script August 9, 2023 9:46 AM

Why would we want legislation written by autocomplete? Who has asked for that? I’d like to see laws against boilerplate legislation mills – I want more local human oversight of lawmaking, not less.

“AIs” are not going to spontaneously create a political party, or anything else. Somebody might program them to do things that look like political activity but how do you foresee them doing something spontaneous? Are they going to build little robot bodies so they can inspect potholes and air quality, or examine libraries’ children’s books for objectionable content about the existence of gay people?

JonKnowsNothing August 9, 2023 10:23 AM

@Clive, All

re: Food Supply Safety v Food Supply Delivery

I split the issue of global food insecurity (not enough to eat) into multiple parts. It gets complex as you dig deeper into the topic. Spoiler Warning: No Fix.

An aspect that gets missed in general, is the economic background of Commodity Markets. Commodity Markets trade in real tangible items in large quantities. It’s similar to how the Stock Markets work except with Stock Markets you get an electronic entry in a ledger (1) and with a Commodity you might end up with the goods dumped on your front lawn.

The idea in Commodity Trading is to build up the middle section, scrape off some profit and make sure the delivery goes to someone else, unless you have a specific use for it.

  • Think train loads of stuff

Most farmers sell their output, through various methods (coop) to a larger buyer. There are some huge companies like ADM and Purina that can handle their own chain from start to finish, but most farmers sell to a Middle Buyer going up to a Commodity Broker.

  • All that UKR grain from 2022 was sold a long time ago to Commodity Brokers
  • All that UKR grain from 2022 is owned by a Commodity Brokerage and sitting in silos in Poland, Hungary and Romania
  • All that UKR grain from 2022 had a Delivery Contract to a destination.
  • Any future profit from the sale of the UKR grain from 2022 is going to the Commodity Brokerage
  • The potential loss of the product through mold, water, rats, etc. is a liability to the Commodity Broker
  • The farmer is not in the picture
  • The destination is not in the picture either
  • It’s about shifting silos of stuff before it rots and there isn’t any profit left
  • It’s also about having empty silos to store the UKR 2023 harvests

When the UKR 2022 harvest was stored in the silos of Europe and not transported to their contract destination, the price soared 400% in some cases. The global markets reacted and the prices skyrocketed. Nothing like having a huge producer like UKR suddenly out of the economic competition. Like the stock market, other players began to fill in the empty delivery slots. Commodities were delivered at a higher price and the folks holding the stuff in Europe saw a great profit potential if they could move their product to destination.

As the market fluctuated, farmers did not get any improved pricing for their harvests. The market fluctuation is a Stock Market game and does not transfer to the farmer. The price the farmer got actually dropped since there were silos of product waiting for the right price point and that caused some outrage from the farmers: No extra profit and the silos were all still filled with 2022 product with no room for 2023 harvest.

So, when there are headlines of shipping being attacked it’s a good idea to consider who is profiting by restricting the transport of goods? Also be aware that there are many ways to move product around the globe and one river is not the only means to get it to a destination. If it’s sitting in silos, there’s an economic reason it’s sitting there and delivery method is not one of them.


1) We used to get a paper Stock Certificate, now it’s an entry in a database.

David Manheim August 15, 2023 5:55 AM

Regarding “Milestone #3: AI-generated political messaging outscores campaign consultant recommendations in poll testing,” I have a single data point that suggests we’re probably long past this by now. While working for Guarding Against Pandemics, which (full disclosure) was funded by Sam Bankman Fried, and is now obviously defunct, I heard from political polling people that this already happened with GPT 3, before the 2022 election, and later heard from the same people with that it got even better with GPT 3.5, though I never saw those results.

Specifically, this was when they were doing message testing for pandemic preparedness investments for Republicans and for Democrats, where the top 5 or so suggestions from GPT3 for each party were tested against 5 or so phrases generated by each of a few people working for the firm – for one party, it had the top message based on the polling, and for the other party, the GPT-generated messages had places 2 and 3. However, from what I recall, the top 3-4 were all reasonably close in terms of sentiment / polling scores – though in both cases there were lots of messages by professional messaging people that did much worse. Given that, it’s not robustly outperforming humans, and this may have been a fluke, but I doubt it, and again, I heard it was getting better with GPT 3.5.

Of course, I suspect these systems not going to get any worse at this, and ever more narrowly segmented messages which can be generated automatically for arbitrary political positions and arbitrary subpopulations or even individuals seem very worrying as political propaganda tools. Given what was happening almost a year ago, I suspect this is already happening at many firms, and they are being used and will be deployed widely in the 2024 election, whether or not anyone admits this.

Leave a comment


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.