How AI Could Write Our Laws

Nearly 90% of the multibillion-dollar federal lobbying apparatus in the United States serves corporate interests. In some cases, the objective of that money is obvious. Google pours millions into lobbying on bills related to antitrust regulation. Big energy companies expect action whenever there is a move to end drilling leases for federal lands, in exchange for the tens of millions they contribute to congressional reelection campaigns.

But lobbying strategies are not always so blunt, and the interests involved are not always so obvious. Consider, for example, a 2013 Massachusetts bill that tried to restrict the commercial use of data collected from K-12 students using services accessed via the internet. The bill appealed to many privacy-conscious education advocates, and appropriately so. But behind the justification of protecting students lay a market-altering policy: the bill was introduced at the behest of Microsoft lobbyists, in an effort to exclude Google Docs from classrooms.

What would happen if such legal-but-sneaky strategies for tilting the rules in favor of one group over another become more widespread and effective? We can see hints of an answer in the remarkable pace at which artificial-intelligence tools for everything from writing to graphic design are being developed and improved. And the unavoidable conclusion is that AI will make lobbying more guileful, and perhaps more successful.

It turns out there is a natural opening for this technology: microlegislation.

“Microlegislation” is a term for small pieces of proposed law that cater—sometimes unexpectedly—to narrow interests. Political scientist Amy McKay coined the term. She studied the 564 amendments to the Affordable Care Act (“Obamacare”) considered by the Senate Finance Committee in 2009, as well as the positions of 866 lobbying groups and their campaign contributions. She documented instances where lobbyist comments—on health-care research, vaccine services, and other provisions—were translated directly into microlegislation in the form of amendments. And she found that those groups’ financial contributions to specific senators on the committee increased the amendments’ chances of passing.

Her finding that lobbying works was no surprise. More important, McKay’s work demonstrated that computer models can predict the likely fate of proposed legislative amendments, as well as the paths by which lobbyists can most effectively secure their desired outcomes. And that turns out to be a critical piece of creating an AI lobbyist.

Lobbying has long been part of the give-and-take among human policymakers and advocates working to balance their competing interests. The danger of microlegislation—a danger greatly exacerbated by AI—is that it can be used in a way that makes it difficult to figure out who the legislation truly benefits.

Another word for a strategy like this is a “hack.” Hacks follow the rules of a system but subvert their intent. Hacking is often associated with computer systems, but the concept is also applicable to social systems like financial markets, tax codes, and legislative processes.

While the idea of monied interests incorporating AI assistive technologies into their lobbying remains hypothetical, specific machine-learning technologies exist today that would enable them to do so. We should expect these techniques to get better and their utilization to grow, just as we’ve seen in so many other domains.

Here’s how it might work.

Crafting an AI microlegislator

To make microlegislation, machine-learning systems must be able to uncover the smallest modification that could be made to a bill or existing law that would make the biggest impact on a narrow interest.

There are three basic challenges involved. First, you must create a policy proposal—small suggested changes to legal text—and anticipate whether or not a human reader would recognize the alteration as substantive. This is important; a change that isn’t detectable is more likely to pass without controversy. Second, you need to do an impact assessment to project the implications of that change for the short- or long-range financial interests of companies. Third, you need a lobbying strategizer to identify what levers of power to pull to get the best proposal into law.

Existing AI tools can tackle all three of these.

The first step, the policy proposal, leverages the core function of generative AI. Large language models, the sort that have been used for general-purpose chatbots such as ChatGPT, can easily be adapted to write like a native in different specialized domains after seeing a relatively small number of examples. This process is called fine-tuning. For example, a model “pre-trained” on a large library of generic text samples from books and the internet can be “fine-tuned” to work effectively on medical literature, computer science papers, and product reviews.

Given this flexibility and capacity for adaptation, a large language model could be fine-tuned to produce draft legislative texts, given a data set of previously offered amendments and the bills they were associated with. Training data is available. At the federal level, it’s provided by the US Government Publishing Office, and there are already tools for downloading and interacting with it. Most other jurisdictions provide similar data feeds, and there are even convenient assemblages of that data.

Meanwhile, large language models like the one underlying ChatGPT are routinely used for summarizing long, complex documents (even laws and computer code) to capture the essential points, and they are optimized to match human expectations. This capability could allow an AI assistant to automatically predict how detectable the true effect of a policy insertion may be to a human reader.

Today, it can take a highly paid team of human lobbyists days or weeks to generate and analyze alternative pieces of microlegislation on behalf of a client. With AI assistance, that could be done instantaneously and cheaply. This opens the door to dramatic increases in the scope of this kind of microlegislating, with a potential to scale across any number of bills in any jurisdiction.

Teaching machines to assess impact

Impact assessment is more complicated. There is a rich series of methods for quantifying the predicted outcome of a decision or policy, and then also optimizing the return under that model. This kind of approach goes by different names in different circles—mathematical programming in management science, utility maximization in economics, and rational design in the life sciences.

To train an AI to do this, we would need to specify some way to calculate the benefit to different parties as a result of a policy choice. That could mean estimating the financial return to different companies under a few different scenarios of taxation or regulation. Economists are skilled at building risk models like this, and companies are already required to formulate and disclose regulatory compliance risk factors to investors. Such a mathematical model could translate directly into a reward function, a grading system that could provide feedback for the model used to create policy proposals and direct the process of training it.

The real challenge in impact assessment for generative AI models would be to parse the textual output of a model like ChatGPT in terms that an economic model could readily use. Automating this would require extracting structured financial information from the draft amendment or any legalese surrounding it. This kind of information extraction, too, is an area where AI has a long history; for example, AI systems have been trained to recognize clinical details in doctors’ notes. Early indications are that large language models are fairly good at recognizing financial information in texts such as investor call transcripts. While it remains an open challenge in the field, they may even be capable of writing out multi-step plans based on descriptions in free text.

Machines as strategists

The last piece of the puzzle is a lobbying strategizer to figure out what actions to take to convince lawmakers to adopt the amendment.

Passing legislation requires a keen understanding of the complex interrelated networks of legislative offices, outside groups, executive agencies, and other stakeholders vying to serve their own interests. Each actor in this network has a baseline perspective and different factors that influence that point of view. For example, a legislator may be moved by seeing an allied stakeholder take a firm position, or by a negative news story, or by a campaign contribution.

It turns out that AI developers are very experienced at modeling these kinds of networks. Machine-learning models for network graphs have been built, refined, improved, and iterated by hundreds of researchers working on incredibly diverse problems: lidar scans used to guide self-driving cars, the chemical functions of molecular structures, the capture of motion in actors’ joints for computer graphics, behaviors in social networks, and more.

In the context of AI-assisted lobbying, political actors like legislators and lobbyists are nodes on a graph, just like users in a social network. Relations between them are graph edges, like social connections. Information can be passed along those edges, like messages sent to a friend or campaign contributions made to a member. AI models can use past examples to learn to estimate how that information changes the network. Calculating the likelihood that a campaign contribution of a given size will flip a legislator’s vote on an amendment is one application.

McKay’s work has already shown us that there are significant, predictable relationships between these actions and the outcomes of legislation, and that the work of discovering those can be automated. Others have shown that graphs of neural network models like those described above can be applied to political systems. The full-scale use of these technologies to guide lobbying strategy is theoretical, but plausible.

Put together, these three components could create an automatic system for generating profitable microlegislation. The policy proposal system would create millions, even billions, of possible amendments. The impact assessor would identify the few that promise to be most profitable to the client. And the lobbying strategy tool would produce a blueprint for getting them passed.

What remains is for human lobbyists to walk the floors of the Capitol or state house, and perhaps supply some cash to grease the wheels. These final two aspects of lobbying—access and financing—cannot be supplied by the AI tools we envision. This suggests that lobbying will continue to primarily benefit those who are already influential and wealthy, and AI assistance will amplify their existing advantages.

The transformative benefit that AI offers to lobbyists and their clients is scale. While individual lobbyists tend to focus on the federal level or a single state, with AI assistance they could more easily infiltrate a large number of state-level (or even local-level) law-making bodies and elections. At that level, where the average cost of a seat is measured in the tens of thousands of dollars instead of millions, a single donor can wield a lot of influence—if automation makes it possible to coordinate lobbying across districts.

How to stop them

When it comes to combating the potentially adverse effects of assistive AI, the first response always seems to be to try to detect whether or not content was AI-generated. We could imagine a defensive AI that detects anomalous lobbyist spending associated with amendments that benefit the contributing group. But by then, the damage might already be done.

In general, methods for detecting the work of AI tend not to keep pace with its ability to generate convincing content. And these strategies won’t be implemented by AIs alone. The lobbyists will still be humans who take the results of an AI microlegislator and further refine the computer’s strategies. These hybrid human-AI systems will not be detectable from their output.

But the good news is: the same strategies that have long been used to combat misbehavior by human lobbyists can still be effective when those lobbyists get an AI assist. We don’t need to reinvent our democracy to stave off the worst risks of AI; we just need to more fully implement long-standing ideals.

First, we should reduce the dependence of legislatures on monolithic, multi-thousand-page omnibus bills voted on under deadline. This style of legislating exploded in the 1980s and 1990s and continues through to the most recent federal budget bill. Notwithstanding their legitimate benefits to the political system, omnibus bills present an obvious and proven vehicle for inserting unnoticed provisions that may later surprise the same legislators who approved them.

The issue is not that individual legislators need more time to read and understand each bill (that isn’t realistic or even necessary). It’s that omnibus bills must pass. There is an imperative to pass a federal budget bill, and so the capacity to push back on individual provisions that may seem deleterious (or just impertinent) to any particular group is small. Bills that are too big to fail are ripe for hacking by microlegislation.

Moreover, the incentive for legislators to introduce microlegislation catering to a narrow interest is greater if the threat of exposure is lower. To strengthen the threat of exposure for misbehaving legislative sponsors, bills should focus more tightly on individual substantive areas and, after the introduction of amendments, allow more time before the committee and floor votes. During this time, we should encourage public review and testimony to provide greater oversight.

Second, we should strengthen disclosure requirements on lobbyists, whether they’re entirely human or AI-assisted. State laws regarding lobbying disclosure are a hodgepodge. North Dakota, for example, only requires lobbying reports to be filed annually, so that by the time a disclosure is made, the policy is likely already decided. A lobbying disclosure scorecard created by Open Secrets, a group researching the influence of money in US politics, tracks nine states that do not even require lobbyists to report their compensation.

Ideally, it would be great for the public to see all communication between lobbyists and legislators, whether it takes the form of a proposed amendment or not. Absent that, let’s give the public the benefit of reviewing what lobbyists are lobbying for—and why. Lobbying is traditionally an activity that happens behind closed doors. Right now, many states reinforce that: they actually exempt testimony delivered publicly to a legislature from being reported as lobbying.

In those jurisdictions, if you reveal your position to the public, you’re no longer lobbying. Let’s do the inverse: require lobbyists to reveal their positions on issues. Some jurisdictions already require a statement of position (a ‘yea’ or ‘nay’) from registered lobbyists. And in most (but not all) states, you could make a public records request regarding meetings held with a state legislator and hope to get something substantive back. But we can expect more—lobbyists could be required to proactively publish, within a few days, a brief summary of what they demanded of policymakers during meetings and why they believe it’s in the general interest.

We can’t rely on corporations to be forthcoming and wholly honest about the reasons behind their lobbying positions. But having them on the record about their intentions would at least provide a baseline for accountability.

Finally, consider the role AI assistive technologies may have on lobbying firms themselves and the labor market for lobbyists. Many observers are rightfully concerned about the possibility of AI replacing or devaluing the human labor it automates. If the automating potential of AI ends up commodifying the work of political strategizing and message development, it may indeed put some professionals on K Street out of work.

But don’t expect that to disrupt the careers of the most astronomically compensated lobbyists: former members Congress and other insiders who have passed through the revolving door. There is no shortage of reform ideas for limiting the ability of government officials turned lobbyists to sell access to their colleagues still in government, and they should be adopted and—equally important—maintained and enforced in successive Congresses and administrations.

None of these solutions are really original, specific to the threats posed by AI, or even predominantly focused on microlegislation—and that’s the point. Good governance should and can be robust to threats from a variety of techniques and actors.

But what makes the risks posed by AI especially pressing now is how fast the field is developing. We expect the scale, strategies, and effectiveness of humans engaged in lobbying to evolve over years and decades. Advancements in AI, meanwhile, seem to be making impressive breakthroughs at a much faster pace—and it’s still accelerating.

The legislative process is a constant struggle between parties trying to control the rules of our society as they are updated, rewritten, and expanded at the federal, state, and local levels. Lobbying is an important tool for balancing various interests through our system. If it’s well-regulated, perhaps lobbying can support policymakers in making equitable decisions on behalf of us all.

This article was co-written with Nathan E. Sanders and originally appeared in MIT Technology Review.

Posted on March 14, 2023 at 12:01 PM34 Comments


anon March 14, 2023 12:16 PM

We don’t need to stop them. All we really need to do is remove the ‘and other matters’ from the titles of bills. As soon as each bill covers only a single topic, this ceases to be that important….except for the part where the legislators don’t read the bills themselves and vote based on what their party chair says, and not their consituents.

llama March 14, 2023 12:21 PM

Having AI write our laws? Great. Let’s also give it more powers over our infrastructure and lives in general. That way it could learn the best way to govern the human society. Surely nothing could go wrong

Clive Robinson March 14, 2023 1:32 PM

@ Bruce, ALL,

Whilst AI could be harmful to the social fabric of society, there is one greater harm we can easily see.

It is that enabler of the 90% or more of current lobbying interest, as you and your co-autgor note,

“… in exchange for the tens of millions they contribute to congressional reelection campaigns.”

Take the money out of politics and lobbying becomes less easy.

But also “limit the number of bills a year” and the number of amendments, making themhave to be proposed and accepted individually atleast three months prior to a vote on a bill.

Oh and put sunset clauses of say 7years into all bills, such that if not passed by more than 2/3rds they lapse.

That will do more to reduce the immediate harms, those who take money up on the hill can do.

Chelsea March 14, 2023 2:18 PM

One rule for them, another for the rest of us.

Of course this law won’t be applied to the tory party themselves – there will be exemptions for national security matters, policing, official secrets, yadda yadda.. But it will be used against everyone else, especially pesky journalists..

Some ex-footballer with an unusually high-functioning brain for the profession decides to criticise the government off-air, this should not have been news (the criticism was well-deserved, if poorly delivered). The news was that he was shut down for it.. Yet the government-appointed Director General (whom is himself subject to criticism for facilitating an £800k loan to Boris Johnson before his appointment…) refuses to comment on whether Lineker would have received the same treatment if he had praised the government… Of course he wouldn’t. As the ‘i’ paper points out this morning, Tory Peer Karen Brady doesn’t get told off by the Beeb for having her Sun column while starring on The Apprentice.. The Beeb is supposed to be impartial, yet the government expects it to act as its state broadcaster. You won’t find the real reason for ditching the European Court of Human Rights explained on the BBC under their current directorship.

And yeah, the ‘i’ article also explains how this farce is being used by the knuckle-dragging right. So that particular incident (leaning on the BBC to shut down a commentator) could have been a deliberate attempt to distract the public away from its plans to restrict their freedoms, whilst at the same time doing further damage to the BBC. Two birds with one stone.

Just as Mr Orwell foresaw, there are different rules for the Party, the Inner Party and the Proles. We criticise China for their freedom-crushing authoritarianism, yet we are implementing the same policies here. The state will control what passes for truth or misinformation.. Denying people a free life based on race is absolutely nothing like what was being discussed in pre-ww2 Germany, and intercepting all communications and outlawing encryption is absolutely nothing like the 1960s Stasi.. right?

It really is nothing to do with Thinking of the Children, it’s all about cementing power by controlling truth and cracking down on dissent. Keeping those rich and powerful in power and riches. That’s pretty much big-C Conservatism in a nutshell, right? Fiefdom for us, serfdom for you.

I am reminded of when I was trying to look up whether hereditary lords and barons pay any inheritance tax on their “substantial landed estates”. (the rest of us pay about 50% and I can’t imagine Lord Muck paying that much on his stately pile…) so I first went to this page and eventually ended up here, only to be told “(This content has been withheld because of exemptions in the Freedom of Information Act 2000)”. WTF? Even the tax manual has sections redacted for the landed gentry, it seems?

JonKnowsNothing March 14, 2023 3:26 PM

@Clive, All

One interesting aspect, soon to be noticed, is the collision of AI LawBot Proposals. It’s not new news that lobbyists pump for laws that are favorable to their client’s cash flow streams to the exclusion of competitor’s cash flows (both competitor client and competitor lobbyist).

iirc(badly) The Ag committee in USA is a powerful committee. Doesn’t get a lot of public notice and rarely any headlines. The focus of the committee in general is to “improve farming” but the reality is to improve “mega farms and mega farm corporations” and not so much the “small farmer” who has become mostly the share cropper of modern Ag.

At the start of the Ethanol Fuel rage, a particular law went into effect about corn grown for the purpose of Ethanol Fuels. It was complex law with lots of twists and turns.

It took a while for these twists and turns to line up the dots and there was only ONE Outfit that benefited from the whole list of Government funding, grants (which don’t have to be repaid) and other sweetheart deals.

So, these LawBots are going to collide with their proposals. Each team’s LawBot will have to digest the other team’s LawBot writings and try to find a preferential path for their own clientele. It’s going to be a spaghetti of conflicting laws.

Only the first few in a particular area will get their laws IN without competition. But after everyone has their own Clippy LawBot Writer on their desktop, it’s going to be VERY INTERESTING.

One aspect in the USA, is that often when a law contains conflicting rules, those rules go On Hold until the courts sort it out. 50 years for ROE, 150-200yrs and counting for Equal Right To Vote. So it may be some time before those laws get sorted.

So new strategies:

1) Get Our LawBot’s proposal accepted
2) Bork the competition’s LawBot’s proposals
3) Keep the competition In Court by creating an infinite set of conflicting rules

It will be like watching ChessBots play each other. If they are good, neither should win. If one wins, the other needs a logic upgrade until it doesn’t lose. If you cannot design a logic upgrade, get a bigger computer, more memory, faster IO, more FLOPs, more PETAFLOPS.

The spanner is that on average, laws change every 18 months. So maybe we will have a new eSport: LawBot Wars. Perhaps people will bet bitcoins on the outcome.

One Common Human March 14, 2023 4:17 PM

Just like “If anything can go wrong, it will”, it is true of AI (Machine Learning) as well. If anything can be misused, it will be misused.

Anonymous-not you March 14, 2023 4:49 PM

Having AI write our laws? Great. Let’s also give it more powers over our infrastructure and lives in general. That way it could learn the best way to govern the human society. Surely nothing could go wrong

Perhaps having A.I. systems review proposed legislation and simplify it would be good start?
Or have A.I. draft legislation without lobbyist involvement, track amendments, and predict the impact for each amendment?
I’d like A.I. to be used to create fair state and US congressional districts that attempt to provide correct representation for as many sub-groups as possible. Right now, my district lines are involved in a lawsuit because 1 political party took advantage and specifically drew the lines so that prior office holders in the other party were forced into a new district or would need to move 1 mile west of their current home to stay in their prior district number new boundaries. It is required that the US Representative actually live in the district they represent here.

lurker March 14, 2023 7:47 PM

When humans hand over their governance to the machines, that’s another step closer to the singularity. It’s no use saying the courts will sort it: apart from the time factor mentioned above, current evidence points to judges trusting the machine rather than the humans who were supposed to be controlling it.

anon March 14, 2023 8:59 PM

The correct answer is contained in the first two sentencs of the essay, and it’s not technical. Not that it would ever be implemented.

Corporate lobbying (private or public) should be illegal, full stop. Companies do not get to vote, rightfully so, so why should they get to influence laws.

Sure, they’ll try to work around it through proxies and by creating large publicity campaigns to try and engage the electorate to act for them, but that adds necessary friction, risks, and cost that provides a much better balance than what most democracies have now.

JPA March 15, 2023 1:19 AM

Reminds me of the old Mad Magazine “Spy vs Spy”

agree 100%

I wonder if machines (or any entity) would be able to assess impact with enough certainty to avoid the potential for significantly harmful consequences to the organization requesting the micro-legislation. The system they are trying to influence has a lot of chaotic characteristics (in the mathematical sense) and small perturbations can have large effects that are mathematically impossible to predict.

I think a small perturbation is more likely to have such unpredictable effects than a large one.

Grima Squeakersen March 15, 2023 8:50 AM

@anon re: “the correct answer” You stop one (very long) step short of the solution, at least for the US. In the US, corporations have long been held by the courts to be legal “persons” that have culpability for their actions, to the exclusion of holding officers and stockholders responsible. By the same doctrine, corporations are also entitled to make political contributions, as contributions are an aspect of political speech, and as such are guaranteed by the First Amendment to the US Constitution. Both aspects of that doctrine are sources of considerable evil. To forbid corporate poltical contributions without abrogating free speech rights for all (which would be an even greater evil) would require dismantling the corporate personage doctrine. I’m skeptical that will ever happen. Even if the political will materialised regarding use of that tactic to reduce or eliminate corporate political contributions, the effects of restoring accountability for corporate actions and consequences to officers and stockholders of corporations (where, imo, it should always have rested) would appear to require a massive restructuring of Wall Street and the rest of the financial system. That would potentially create as much chaos as an outright revolution overthrowing the current form of government, and is equally unlikely to occur.

Duckface March 15, 2023 11:27 AM

None of this addresses the real problem.

Taking donations for law amendments is bribing, and should be treated as such by prosecuting everyone involved.

Petre Peter March 15, 2023 12:43 PM

If machines start writing laws, then “we will have laws that are as upgradable as our computers. “

ResearcherZero March 16, 2023 6:31 AM


“What precisely caused the observed reversal of the long-term trend around 1980 remains perhaps even more difficult to pinpoint. However, according to the authors there could be a connection to tensions arising from changes in economic policies since the early 1980s, which may have been defended on rational arguments but the benefits of which were not equally distributed.”


ResearcherZero March 16, 2023 7:59 AM

Transparency would definitely have many benefits. It may allow some incite into the arguments of those with a large amount of power, who seek a cost performance benefit over those with substantially less power.


In follow-up cases in Northern Territory courts, it was established that the Department of Housing chief executive, was required to provide habitable housing, which is “humane and reasonably comfortable”.

Lawyers for Ms Young and the other tenant in the case have said they believe the case will have ramifications for renters across Australia.

“We live in a hot area and have no air conditioning.”

The class action alleges it was unconscionable for the NT chief executive of housing to fail to repair houses, fail to reduce rent where houses were not in good repair and fail to properly explain tenancy agreements, in circumstances where tenants had no other option for housing.

Mr Kelly said the findings now set a standard that applied to all remote communities going forward.

“All landlords across the NT [now] need to provide housing that provides a reasonable level of comfort and in all the circumstances is humane.”

Court documents show lawyers representing the NT Government will argue that this part of Justice Blokland’s ruling was wrong, with one ground of the appeal reading:

“The learned judge erred in construing section 48(1)(a) of the Residential Tenancies Act by holding that “habitable” requires an overall assessment of humaneness, suitability and reasonable comfort of the premises judged against contemporary standards.”

Building houses for humans that are “habitable” does sound quite insane!

Carbon-based error or biological interface error? Perhaps ‘machines’ are already involved in the legal process, and have long demonstrated they are living among us?

And could AI further compound the problem?

When the variables are the values of experimental measurements they have uncertainties due to measurement limitations (e.g., instrument precision) which propagate due to the combination of variables in the function.

This page shows how uncertainty in a measured quantity will propagate through a mathematical expression involving that quantity.



Phillip March 16, 2023 4:35 PM

Given the present legislative strategy of hiding any nuance with the felling of too many trees, this is a believable step forward.

With nobody realizing anything, let us step up the call to cease unending debate over whether anybody has sufficient cause to speed-read in the deed-of-the-night?

All right, I will admit anything is positive. Imagine the political hit piece, when so-and-so accuses so-and-so of illiteracy: “How will any of it ever make a gosh-darn lick of sense?” Au contraire: “Notice the accent mark, my word!”

Will it become one’s civic duty to watch a little late-night comedy?

Phillip March 16, 2023 5:39 PM


I looked at your link about compounding error, vis-a-vis, way too much measurement-assuming.

Hmm. Not long ago, I was fascinated to learn how a usual probability model used in Quantum Mechanics (QC) differed from a usual classical probability model. QC does build upon classical with an added term, or expression, or what one might call it.

I understand the difference in the QC model owes to Heisenberg’s Uncertainty Principle, or something to this effect. Anyway, I came about the topic while watching much of a YT series with content from Professor Artur Ekert. Funny thing is, I was dumbfounded. I was certain I discovered why I struggled with classical probability. I did not know. While it answers the mail, this doesn’t entirely make sense, I will admit.

ResearcherZero March 17, 2023 2:20 AM


Richard Feynman’s videos at Caltech are probably some of the best, and there are a lot of them. As Feynman points out often, it is dumbfounding. But he does do a good job at explaining why probability is the way it is. And many of the lectures are fairly easy to follow, until at least you start to get to the later lectures, where he points out regularly that it doesn’t seem to make any sense.


ResearcherZero March 17, 2023 2:22 AM

At the moment, the only information available to the public shows the identity of the lobbyists and their clients, but not their meetings or activities with public representatives. The compliance system relies on self-regulation.

The Attorney-General’s Department did not implement the recommendation from Auditor-General Report No.27 of 2017-18, Management of the Australian Government’s Register of Lobbyists.

In early 2018, the Australian National Audit Office made a raft of recommendations to government to improve the system.

A follow-up audit … found the Attorney General’s Department (AGD) had largely failed to act on the recommendations in the more than two years since.

“no method” available to the department to work out whether lobbyists were being “transparent about whose interests they are representing when meeting with a government official”

The department also had “no power” to check whether government officials were reporting known breaches of the lobbying rules as required, or that they were checking to ensure the lobbyists they were meeting with had listed themselves on the lobbyist register.

“AGD has no means to verify that government representatives are checking the register and meeting only with registered lobbyists and with former government representatives only outside the prohibition period,” the audit found.

“As at 31 January 2020, 29 former government representative lobbyists had ceased employment as a government representative within the previous 12 months and were therefore prohibited from lobbying on certain topics. In the period between May 2019 and January 2020, AGD notified six of these lobbyists (via their employer) about prohibitions,” it found. “In three instances, this was more than 48 days after the relevant record had been created in the register.”

This applies for 18 months after they have left their roles, and 12 months for senior public servants and military figures.

The register doesn’t show the past government links of at least 75 lobbyists, including former state premiers, party assistant secretaries and chiefs of staff to federal opposition leaders. The definition of a lobbyist is also extremely narrow.

ResearcherZero March 17, 2023 2:30 AM

Bruce and Nathan’s article is both timely and excellent. Especially the great deal of thought and research providing suggestions of how to tackle the issue.

Lobbying spending continues to grow at both the state and federal levels, highlighting the importance of shining light on these efforts to impact public policy. In fact, 2022 first quarter disclosures indicate spending this year may set a new record.

“an increasingly concerted effort by national players organizing lobbying activities across state and federal jurisdictions”


To highlight the text again:

“We don’t need to reinvent our democracy to stave off the worst risks of AI; we just need to more fully implement long-standing ideals.”

“…Moreover, the incentive for legislators to introduce microlegislation catering to a narrow interest is greater if the threat of exposure is lower. To strengthen the threat of exposure for misbehaving legislative sponsors, bills should focus more tightly on individual substantive areas and, after the introduction of amendments, allow more time before the committee and floor votes. During this time, we should encourage public review and testimony to provide greater oversight.”

“Some jurisdictions already require a statement of position (a ‘yea’ or ‘nay’) from registered lobbyists.”

“But we can expect more—lobbyists could be required to proactively publish, within a few days, a brief summary of what they demanded of policymakers during meetings and why they believe it’s in the general interest.”

“We can’t rely on corporations to be forthcoming and wholly honest about the reasons behind their lobbying positions. But having them on the record about their intentions would at least provide a baseline for accountability.”

Winter March 17, 2023 3:37 AM


As Feynman points out often, it is dumbfounding. But he does do a good job at explaining why probability is the way it is.

I understood that the problem has its roots in the incompatibility between laws of nature that are deterministic (all dynamics is unitary) on the one hand and probability requiring non-deterministic behavior on the other hand.

Solutions to this incompatibility basically rely on making information invisible (coarse graining in entropy) or inaccessible (too complex to calculate, eg, quantum gravity).

So we are in a bind, Physics is deterministic, but observations show “probability” (noise) is everywhere. Which means that physics tells us there is no true randomness and experience, and thermodynamics, tells us otherwise.

ResearcherZero March 17, 2023 5:31 AM


There are a few clues, and a lot of research opening up in new fields. Perhaps some of the big fancy equipment will give us some answers, but still a lot of discoveries have come from experimentation in the lab, and working out why some of the results are unexpected and strange.

“The presence of dielectric material affects other electrical phenomena. The effects of the dielectric on electrical phenomena are described on a large, or macroscopic, scale by employing such concepts as dielectric constant, permittivity, and electric polarization.”

There is some interesting new research in many other areas.

A simple catch: Fluctuations enable hydrodynamic trapping of microrollers by obstacles

“Driscoll and her team found the hydrodynamics of the fluid inside the sample chamber created stagnant areas. In other words, the spinning microroller caused the fluid to flow in the chamber. But the flows also created pockets—including one directly behind the obstacle—where the fluid remained still and unflowing. When the particle entered the stagnant area, it stopped moving and became stuck.”

“But to reach the stagnant area, the particle had to perform a baffling U-turn. After moving past the obstacle, the microroller curved around it, becoming stuck to its backside. Driscoll found that random motions (called Brownian motion) of the molecules within the fluid “kicked” the microroller into the stagnant area.”

“Tiny materials are subject to Brownian fluctuations. The fluid is not actually a continuum but is composed of individual, little molecules. Those molecules are constantly ramming into the particle at random orientations. If the particle is small enough, these collisions can move it.”

Clive Robinson March 17, 2023 6:08 AM

@ Winter, ResearcherZero, ALL,

Re : What you can not measure.

“So we are in a bind, Physics is deterministic, but observations show “probability” (noise) is everywhere. Which means that physics tells us there is no true randomness and experience, and thermodynamics, tells us otherwise.”

The “determanistic” and “probability” are both correct.

And noise is not “random” it’s just easier to treat it as so.

Likewise “probability” deals with averages of multiple entities moving with a complexity we can not measure. However the “bulk” effect is very predictable.

To see why think about “Brownian Motion” and the various gas and other fluid laws.

As you supply energy things happen to individual particles that may or may not be discrete (the jury is still out). However we know very accurately that the working fluid will expand and why, and this alows us to determine certain “bulk effects” under gravity etc.

However as any school science student who has watched the experiment with one flask of purple water at the bottom connected by two glass tubes to an upper clear flask of water will know the bulk effect on heating is,

“The purple water moves up and the clear water moves down untill they are equally mixed”

Exactly as predicted by bulk effects mathmatics based on the averaging probability effect.

However what the student can not explain nor can better qualified scientists is how the purple water moves in the clear water and vice verser for any given run of the experiment.

In those brief few moments of the purple jet of water mixing in the clear water is more complexity than we will ever be able to fully calculate or even comprehend in a life time. And like snow flakes every run of the experiment will be unique in those moments.

We can describe mathmatically very deterministically what happens with each molecule yet we can not measure them in the experiment…

In those moments we see one of the most important things to know about the nature of “noise”. That is how things average over time in ways to complex to measure.

That is each molecule behaves “determanisticaly” but with millions of others it intersects many times so it’s path is what most would call chaotic, and the effect when seen in bulk is what most would call random.

One of the fundemental tenents of science is reversability for that to be possible each entity must have “memory” of every interaction it has ever had, in what order and when.

The implication of this is the entities state is the sum of all those interactions. As such we can not examine that sum to find out those individual events, we can only unwind each interaction one by one. Which we realy can not do for the obvious reason we are not omniscient, nor can we look back in unrecorded time.

So what we can not measure, and can not know, can only be treated by averaged calculations of entities in bulk. It works, to a very high degree with sufficient entities, but there is always imprecision a shadow of what we can not know, and this we call “noise”. It is the sum of all the individual signals both negative and positive, that average to some value, but it is for ever changing as the entities which form the “working fluid” move under the influance of forces.

Have you ever wondered what would happen if we did the same purple water experiment in “free fall” or what most would call “in space”?

Well the answer is complex, but a simpler experiment will give you a starting point. You might have heard that “a candle will not burn in space” or similar about a match. There is even a video of such an experiment.

The flame starts and very quickly dies. The reason is that the change in density of the heated gas from combustion does not rise, as it’s not in a gravity gradient. So there is no upwards movment of the hot burnt gasses drawing in oxygen from beneath the flame, so the combustion is starved and goes out.

Now all you have to do is think how that would change an experiment designed to work in a gravity gradient 😉

Winter March 17, 2023 6:09 AM


There are a few clues, and a lot of research opening up in new fields.

Re: True randomness

The question is, what do you want from “random signals”?

If you want True Randomness, you want signals that have no cause. That is, it is not possible to reconstruct the initial conditions from the end conditions, ie, the random signal. Formulated this way, this is most definitely impossible using currently known physical laws. Even if you burn a book, physics leaves open the possibility that you reconstruct (part of) the book from the smoke, soot, and heat.

If, instead, you want thermodynamic randomness, you have to specify the coarse graining cut-off where you specify what microscopic states you consider invisible. But there is always the possibility that some agent is able to peer into some of the microscopic states and thus is able to retrace (part of) your random signal.

vas pup March 17, 2023 4:23 PM

@Bruce said: “The issue is not that individual legislators need more time to read and understand each bill (that isn’t realistic or even necessary). It’s that omnibus bills must pass. There is an imperative to pass a federal budget bill, and so the capacity to push back on individual provisions that may seem deleterious (or just impertinent) to any particular group is small. Bills that are too big to fail are ripe for hacking by microlegislation.”

When new speaker was elected one of the requirements by internal opposition group was: “No more omnibus acts. One subject – one act”. That is very reasonable and useful suggestion to prevent possibility of not reading the act by legislators before voting. Bill should be provided for reading no less than 72 hours before vote.

Phillip March 17, 2023 9:52 PM

@vas pup:

Since our legislative branch is used to punting any war resolution over to the executive, 72 hours is something I can meet you with.

Phillip March 17, 2023 9:54 PM

@vas pup:

By the way, I realize the executive has, like ten days or something. My only motivation is to get your 72-hour idea into practice, what with the problem. Thank you for reminding me, what Bruce wrote.

ResearcherZero March 17, 2023 10:43 PM

@Phillip @Winter

How will a ball behave inside a roulette wheel? You could measure it’s mass and the forces exerted on it. (though they do not allow scientific equipment inside a casino)

If you can make more precise measurements of particles, improved predictions can be made about their properties.

“The field as a whole is entering its precision era, moving beyond just slamming particles into each other to see if they throw off new subatomic bits and adopting meticulous techniques to probe their properties.”

How do you measure something so small?

The precise charge of the electron is already known.

ResearcherZero March 18, 2023 12:49 AM

There are some models for better regulation. In 2018, for example, California passed legislation mandating that chatbots disclose when there isn’t a human on the other side of the conversation. But there is no pending legislation in Congress that demands companies offer a human point of contact.

“It’s a major abdication of corporate responsibility.”

A survey by OnePoll in 2021 found that more than two-thirds of respondents ranked speaking to a human representative as one of their preferred methods of interacting with a company, while 55 percent identified the ability to reach a human as the most important attribute a customer service department can possess.

Nick Levinson March 18, 2023 4:04 PM

@ALL, @anon, @vas pup, @Clive Robinson, & @Duckface:


Private laws have been passed by Congress for decades and I think various state legislatures and likely other legislatures have been doing similar things for long. This is in the U.S. I haven’t researched this level of law in years. The private laws are published (e.g., in U.S. Statutes at Large), but not codified (e.g., in U.S. Code), and some provisions of public laws are also not codified, so they’re harder to find by subject, being mainly chronological. Often they’re to give a nationalization or immigration benefit to one named individual, but they can be for anything. The general rule for most codifications I’ve seen is that they’re for laws that are statutes or statute provisions and that are of a general and permanent nature, all criteria having to be satisfied, “general” meaning applicable to enough people and institutions to be worth codifying rather than just handing to a few (one not codified was on local water rights and involved a nearby community and an Indian tribe) and “permanent” apparently meaning longer than a year, thus an annual budget is not considered permanent for the purpose. One N.Y. Times reporter uncovered a tax law provision that applied only to any company that had been formed in two jurisdictions on a certain date; only one company was known to qualify.

I skimmed the article: very interesting and of concern.

I don’t have much confidence in AI in general or for legislating. I’m sure AI can vastly multiply the number of bills drafted and available for introduction, but I’m not sure it can much increase the number introduced or passed, the bottleneck being people. To widen the neck likely requires that most lobbyists but also most legislative offices each have their own AI systems adapted to their own purposes, i.e., not the same AI for everyone.

If lobbyists grow massively in numbers, with or without AI, then for a lobbyist to be persuasive they’ll usually have to send bigger campaign contributions, and so will others, adding expense through lobbyists, so those who give the same amount will lose out. This will grow the labor pool of those who’d like to be a legislator or work for one and, I think, strengthen their parties.

@anon & @vas pup:

Limiting a bill to a single subject would, I think, lead to more litigation, which would be over whether a law was validly enacted, perhaps years later. Some litigation is for good purposes but I think this would be wasteful.

There are too many bills to expect every legislator or even their staffs to read them all and to understand all the subjects each bill entails. Legislators and their staffs are at least paid to spend time legislating, whereas relatively few constituents even make enough of a living to have time to devote to doing that same work, so they don’t weigh in.

@Clive Robinson:

To take the big money out of politics, you may have to take the big power out of politics first. To take it out of one political venue without moving it to one or more other political venues is probably impossible.

Reducing the ability to legislate, such as by universal sunsetting that would usually force replacement bills to be considered, limiting bills and amendments by quantity so they’d be longer and more complex, adding waiting periods for amendments when sometimes speed is indeed appropriate, and supermajorities when that makes passage less likely (at least in the U..S.) and I wonder if that’s a good idea (it is for some subjects but post-sunset is not subject-related), might make legislatures even less capable than they are now, and the idea of less government is less popular than its proponents wish.


Prosecuting for bribery requires that how money comes into politics be defined as bribery. Most of it isn’t, and a criminal law has to be rather specific about what’s illegal in order to inform people about what’s forbidden. It doesn’t have to be narrow or list every circumstance but it has to be specific. It’s also rare for a law enforcement agency that gets its funding from the people it would consider enforcing against to follow through, although it can happen, as in Abscam (that source omits one Senator who refused to be bribed) (subsequent limits have been reported). One legislator (I forgot who) said that a large donation does not buy a legislative outcome but it does pay for her time to listen to the lobbyist’s idea for (say) 15 minutes. One local candidate called people who were already major donors; thus, they talked; she often was briefed in advance on a donor’s major interest/s and a donor might ask her about something (one asked which three agencies she would close and apparently liked her answer and so gave again).

ResearcherZero March 19, 2023 3:31 AM

“State Of Unholy Matrimony – When Govt Merged With Business”


On the day after the 1983 state election, the incoming premier Brian Burke met with Laurie Connell, who was to become a regular financial adviser, and others, to whom he announced that he wanted to be involved with people in the local business community, and that the new government would set up the Western Australian Development Corporation.

In his role of state Treasurer, Burke had the statutory capacity to direct the financial affairs of substantial corporations including the State Government Insurance Commission (SGIC) and the Superannuation Board, later renamed the Government Employees’ Superannuation Board (GESB), which was given sweeping new powers in 1987, enabling its extensive funds to be used for virtually any purpose approved by the Treasurer.



…a former detective in charge of the case has fled the country after admitting that he used lies, beatings and fabricated evidence to build a case against three brothers convicted of the million-dollar theft.

“This is where you die you little f***er. You’re on another planet, no one knows you’re here. As far as they’re concerned, you could be dead.”

Detectives, led by Don Hancock, had lied at their trial in the District Court, had fabricated confessions by all three, and had planted the damning fingerprint.

The corrupt former police officer at the centre of the Mickelberg-Perth mint swindle affair has been sent to a secure ward of a psychiatric hospital.


Unconventional gold transportation…

Strangely, the two other corrupt cops who accompanied Hancock and Lewandowski everywhere, are not mentioned in any of the stories about the Perth Mint.

ResearcherZero March 19, 2023 3:40 AM

The Premier Alan Carpenter wants to put the lobbying controversy behind him.

“We can leave these characters behind where they belong in the dust bin of history,” he said.

Apparently not.

Leave a comment


Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via

Sidebar photo of Bruce Schneier by Joe MacInnis.