The bill for the party
Let us summarise the bill, for those just arriving. A law born twenty-five years ago to transform the culture of business has produced a single textbook case, which lasted twenty years; a handful of acquittals on the merits for suitability of the model; no standard of good organisation; a suitability assessment that remains a retrospective lottery; penalties the OECD judges inadequate to their purpose; no database and no public statistics. Around this void has grown the richest market in Italian criminal law: models costing up to sixty thousand euros, thousands of supervisory seats with neither a duty nor a liability, an elite of a few dozen names celebrated by the rankings and a crowd of thousands of experts whom nobody counts. In the first two parts we described the product and the market; in this one we look ahead, because in recent months converging signals have matured that could bring the party to a close — or relaunch it in a worse form. From the State has come an organic reform of the decree, delivered to the Minister of Justice. From the courts have come an opening towards standards and the first judgments that describe the machine for what it is. From technology has come generative artificial intelligence, which has learnt to imitate to perfection precisely the product on which the party rests: the paper.
The State's confession: the reform
Let us begin with the State, because the most merciless document on Legislative Decree 231/2001 was written by the State itself. Reforming the decree has been discussed, in conferences and bills, for at least fifteen years, without anything ever reaching port. This time is different: in February 2024 the Ministry of Justice set up a technical committee for the organic revision of the decree, and it did so two years after the failing grade from the OECD, which had asked Italy in writing to act on penalties and extinction periods. Called to coordinate the committee was Giorgio Fidelbo, a section president of the Court of Cassation — the man who presided over the panel in the last Impregilo judgment of 2022, the most important ever written on the subject. The man who had signed the model judgment was tasked with rewriting the law that that judgment had not sufficed to straighten out: this alone, by itself, is a verdict. The committee worked for almost two years, heard Confindustria, Assonime, the association of members of Supervisory Bodies (organismi di vigilanza), the consultants, and delivered a final report with draft articles, made public in early 2026.
The heart of the proposal is a dogmatic surrender disguised as refinement. Organisational fault becomes a constitutive element of the offence, for the crimes of senior management as for those of subordinates, and the burden of proving it passes entirely to the prosecution: the reversal of the burden of proof that for a quarter of a century had been the distinctive trait — and the commercial terror — of the system falls away. Around this core, a coherent set of accessories: rewards for the entity that cooperates and reorganises itself, expanded negotiated justice, a ground for extinguishing the offence tied to a corporate reorganisation proposal, adjustments to the extinction periods the OECD censured. Every single choice is defensible, and some are sacrosanct. But look at the whole through the eyes of the first two parts. In a system in which the public prosecutor has neither instruments for measuring organisation, nor statistics, nor organisational training, shifting onto him the entire burden of proving organisational fault means, in practice, guaranteeing even fewer convictions than the — already almost non-existent — convictions of today. In a system devoid of any metric of effectiveness, rewards for the entity that reorganises itself mean discounts in exchange for new paper: the reorganisation proposal will be written by the same consultants who had written the model that had not prevented the crime, and nobody will measure that one either. And there is a detail worth the whole: among the proposed innovations is the omission, from the certificate of the register of sanctions, of the decisions that settle the offence by way of reward. In a system already opaque — no database, unpublished decisions on the merits, invisible dismissals — the only certain reform of transparency moves in the direction of more opacity.
As for the political paternity of the ideas, no conspiracy theorising is needed: it is enough to read. Confindustria, in its position paper of April 2025, defined the rules in force as a strongly repressive instrument with uncertain boundaries, and asked for rewards, negotiated justice, and recognition of trade-association codes as a parameter for the models; the summary of the proceedings circulated by the Confindustria system claims, in its opening pages, the connections between the committee's proposals and the association's demands. The regulated party wrote its own rule, in broad daylight, with the ministry's stamp. In parallel runs a parliamentary bill that would make the Supervisory Body collegiate, with members enrolled in professional registers identified by decree: politics' answer to the guardian who does not work is to establish a register of guardians — consolidating the profession, not the function.
But the most eloquent document is the list of silences. In the text made public there is no database of decisions; no ministerial statistical survey of proceedings and outcomes; no public parameters for the suitability of models; no intervention on the duties and liability of the Supervisory Body; no organisational training for those who will have to judge. For a law whose problem, declared for twenty-five years, is that nobody knows what makes a model suitable and nobody counts what the law produces, a reform that redefines fault but produces neither parameters nor numbers treats the symptom that irritates the defendant, not the disease that afflicts the system. Any engineer, before redesigning a machine, would demand the data on how it runs; here the machine is being redesigned in the dark, on the instructions of those who must submit to it.
The courts' opening: the purchased standard
The second signal comes from the case law, and it is more ambiguous. In September 2025 the Court of Cassation held that, for models in the field of workplace safety, certified conformity to recognised international standards can count as a rebuttable presumption of adequacy. It is the first time, in a quarter of a century, that the court of last instance has pointed companies to a parameter verifiable before the disaster: the direction — predictability, standards, presumptions — is the one this series has been calling for since the first part. But look at the mechanics: the assessment migrates from the judge to the private certifier, who is chosen and paid by the party being certified, who renews the engagement periodically, and whose audit is itself a market service.
The precedent already exists, and it is the most instructive of all. Italy is the world's leading country for ISO 37001 anti-corruption certifications — a multiple of any other jurisdiction — and in the very years in which it was collecting its record haul of stamps, the OECD was writing that its repression of foreign bribery produces few convictions and penalties that are neither effective nor dissuasive. The stamp has not produced integrity: it has produced stamps. If the presumption of adequacy is extended without a public audit of the certifiers, the paper industry will simply have found its final stamp: the photocopied model, certified in photocopy, presumed suitable by law. The standard is the right road; handing it to the market it is supposed to discipline is the sure way to travel it backwards.
The impatience of the courtrooms
There is then a signal less conspicuous than the reform and the technology, and it comes from the courtrooms. We documented it in the second part and here it is enough to recall it. In Milan, in the trial over the Monte dei Paschi derivatives, the court described a Supervisory Body that had remained "an inert onlooker" before the information flows passing under its eyes; in Venice, in the trial over Popolare di Vicenza, the Court of Appeal spoke of "osmosis" between the controllers and the controlled top management. These are judgments that do not confine themselves to saying that the model did not work: they begin to describe, with perceptible irritation, the real workings of the machine — the perfect paper, the ceremonial supervision. And in 2024, again in Milan, came the second acquittal on the merits for suitability of the model in over twenty years: the second. Anyone who wants to read the signals finds them aligned in the same direction: a body of case law that has stopped believing in the paper, and that has not yet been given the tools to judge anything else.
Because this is the point: impatience without tools produces admonitions, not method. The judge who writes "an inert onlooker" is making, in the negative, the only substantive assessment possible — looking at what the body did, not at what the rulebook assigned to it. But to make it in the positive, to say when supervision is serious, when a flow is real, when an incentive is coherent, requires the parameters, the data and the training that no reform provides. Until then every judgment will remain an episode, and the industry will go on reading even the admonitions as insurable risks: after every conviction the update is sold, after every admonition the training course on the admonition.
The foreign mirror
Comparison with other jurisdictions removes every alibi, and it is worth making it on the point the conferences avoid: who verifies, by what criteria, published where. In the United States the Department of Justice has for years published and updated an operational document — the Evaluation of Corporate Compliance Programs — which lists the questions the prosecutor must put to the company: is the programme well designed? is it endowed with resources and real powers? and above all: does it work in practice? Concrete criteria follow: the data to which the control function has access, the tests carried out, the lessons drawn from failures, the incentives and the careers of those who report and of those who cover up. One may debate how far practice lives up to the document; but the document exists, it is public, and it is written by the prosecution: the American company knows in advance what it will be judged on. The Italian one finds out from the court-appointed expert, after the disaster.
In the United Kingdom, negotiated-justice agreements with companies must be approved by a judge, and the approval decisions are published with extensive reasoning explaining what the company had and had not done: negotiated justice there produces readable precedents, not invisible dismissals. France, with the Sapin II law of 2016, took the step that in Italy has never even been discussed: a public agency — the Agence française anticorruption — which sets the requirements for prevention programmes, enters companies, inspects them and can refer them to a sanctions commission. The controller of the programmes, in France, is the State. In Italy the market sells control to itself: the consultant writes the model, a body chosen by the controlled party supervises it, a certifier paid by the certified party attests it.
Germany, which has no general law on the punishment of legal entities, tried to give itself one in 2020 and saw the bill sink, amid the resistance of the business associations, at the end of the parliamentary term: no law, and at least no party. Italy chose the third way, the worst: the law is there, so is the party, the controls are not. Twenty-five years of conferences have compared Decree 231 with everything — American law, English law, French law — but almost never on the point that matters. It is not a technical detail: it is the entire difference between a system of prevention and a market for documents.
The discipline that comes from abroad
There is a corollary of the foreign mirror that the market knows very well and never says out loud: for the Italian companies that matter, the real discipline comes from abroad. The listed multinationals, the banks, anyone who touches the American or the British market has lived for years under laws with extraterritorial reach — the Foreign Corrupt Practices Act, the UK Bribery Act — and under regulators who genuinely patrol: and for those laws it builds programmes with resources, tests, internal investigations, because it knows the check may come. The same companies, in Italy, also buy the 231 model: but as a local formality, a rite to place alongside the real programme. Elite Italian compliance is thus often double: a substantive one, written in English, for the foreign prosecutor; a ceremonial one, written in Italian, for the remote eventuality of a national judge. The domestic paper market lives in the gap between the two.
The pressure that can end the party, for that matter, also speaks a foreign language. The OECD has placed Italy under follow-up to its Phase 4 evaluation, and the subsequent reports will again count convictions and penalties; European legislative techniques — whistleblowing yesterday, sustainability reporting today, the anti-corruption directive under discussion — impose duties more measurable than the domestic ones; and the large institutional investors have begun to ask their investee companies, in their engagement and voting criteria, for evidence of effective compliance and not attestations. None of this is decisive; but it explains a paradox that anyone watching from abroad notices at once: the country that produced the most theorised organisational model in Europe is the same one in which the effectiveness of prevention is demanded, when it is demanded at all, by the authorities and markets of other countries.
The economics of fear
Before coming to the technology, one question must be asked, because it is the key to everything else: why has the market never corrected itself? In any other sector, a product that never proves it works loses its price. Not here, and the reason is that the product is not prevention: it is procedural insurance, and insurance is judged only when the claim comes. The figures in the second part say the claim barely exists: a few hundred judgments on the merits in twenty-five years, against more than a million companies potentially subject to the decree; in the most active district in Italy, a few dozen convictions of entities in seven years; for foreign bribery, twenty-two convictions in twelve years by the OECD's count. The probability that the model will ever be read by a judge is, for the average company, close to zero. The rational buyer, then, is not buying prevention — which is invisible — nor even, at bottom, protection: he is buying a talisman against a remote risk, and for a talisman what counts is the signature of the seller, not the content. It is the only market in which the price forms entirely on the seller's reputation, because the quality of the product is never measured by anyone.
And when the claim comes, the insurance almost never pays: in the trials, the models — where they exist — are almost always judged unsuitable, as the only two acquittals for suitability in over twenty years demonstrate better than any comment. An insurance policy on which claims almost never arise and which, when they do, almost never pays out would have an unpleasant name in any regulated market; here it is called best practice, and the premium is justified by the rite, not by the payout. This is why the party could not end from within: neither the seller nor the buyer has an interest in measuring, and those who ought to impose measurement — legislator and judge — have never done so. An equilibrium of this kind can be broken only from outside, by one of three forces: a law that changes the incentives, a judge who changes the yardstick, or a collapse in the cost of production that changes the prices. The first two, as we have seen, hesitate. The third has just arrived.
The machine that writes for free
Generative artificial intelligence has entered the Italian legal market through the front door: the majority of the large firms have adopted it or are trialling it, the technology spending of professional firms is growing at double-digit rates, and the 231 segment — made of serial documents, repetitive updates, standard formats — is the most exposed of all. The signals are already on the market, and we do not hide them out of delicacy: collections of prompts for drafting 231 models are sold on subscription for a few dozen euros a month; the management platforms for Supervisory Bodies incorporate automated assistants; even the sector's journal of reference has launched its own answer engine trained on its own archive. The paper market has begun to automate itself.
The point is not technological: it is economic, and it is brutal. For twenty-five years the price of the model — five, twenty, sixty thousand euros — was justified by the expert's time: the mapping, the interviews, the writing, the updating. A machine today produces in an hour a document indistinguishable, in form, from that of the prestigious firm: general part, special parts, risk matrix, code of ethics, flows to the Supervisory Body. The marginal cost of the paper has fallen to zero. In the short term the supplier pockets the differential: same fee, a tenth of the time — and the sector's silence on the subject explains itself. In the medium term the buyer learns: he asks for the discount, brings the work in-house, or buys the subscription for a few dozen euros. The information asymmetry that held the market up — the client does not know what he is buying, nor what it costs to produce — turns for the first time against the seller. A market built on the price of paper does not survive free paper: it can only change product, or change story.
Where the rent retreats
When the price of paper collapses, the rent does not evaporate: it retreats into the rooms the machine cannot occupy. The first room is the seat on the Supervisory Body. The model is sold once; the body is sold every year: fees from a few thousand to tens of thousands of euros per entity, for thousands of entities, with no measurable duties and no liability — the perfect rent, because it is recurring. An artificial intelligence can write the body's rulebook in a minute; it cannot sit at the quarterly meeting, sign the minutes, collect the attendance fee. No surprise, then, that while the value of the paper collapses, the parliamentary bill in circulation concerns not the paper but the seat: the register of members. When the product is devalued, access to the rent is fortified. The second room is the story: if the criminal fear is objectively low — few convictions, modest penalties, time limits that extinguish — the seller moves the threat elsewhere: reputation, ratings, public tenders, the supply chain, sustainability. Already today the conferences announce the integration of the 231 model into ESG systems: the paper changes its letterhead, the fee does not.
The counter-proof lies in the price lists. While the technology was driving the cost of producing the document to zero, we are not aware of a single announcement of a reduction in the prices of models; what we do find are new products — the integrated model, the sustainable model, artificial intelligence governance. It is the typical behaviour of a market without price competition facing a cost shock: the price list is not lowered, the menu is widened. It will work until the buyer learns to ask what he is paying for; and this series, among other things, would like to shorten the learning time.
Self-reference, declared
At this point we must say how this series was built, because the argument is worth more than any abstract demonstration. The investigations you have read — the case law surveyed, the market inventoried, the Supervisory Body X-rayed — were prepared with commercial artificial intelligence tools, within anyone's reach: searches across hundreds of open sources, cross-checking of references, reconstruction of repealed rules, censuses of conferences and fee scales. Work that would once have required months of a researcher's time was compressed into days; and every statement was then checked back against its source, because the machine makes mistakes, and the points we could not verify we have declared as such. Anyone can draw the commercial implication: if an independent collective can do this to criticise the market, any in-house legal department can do it to stop paying the market's rent. But the deeper implication is the other one: what the machine did not do, and cannot do, is precisely what in this field would always have been the only thing worthy of a fee — taking responsibility for a judgment, knowing an organisation from the inside, looking a board of directors in the eye and telling it what it does not want to hear. The paper was not worth sixty thousand euros yesterday and is not worth zero today: its value has always been its current one, zero. Artificial intelligence did not reduce it to zero; it revealed it.
The next party is already laid out
Here it is worth widening the frame, because Decree 231 was not an episode: it was a prototype. Observe the sequence of compliance markets in Italy. In 2001, the liability of entities, with the model and the Supervisory Body. In 2018, data protection, with the European regulation: records of processing, impact assessments, data protection officers appointed in their thousands, and a wave of photocopied consultancy that practitioners themselves have learnt to call façade compliance. In 2023, whistleblowing: mandatory reporting channels for tens of thousands of companies and, within months, an endless supply of platforms and subscription packages. In 2024, sustainability: European reporting generated reports of hundreds of pages and a new army of consultants, even before the Union, frightened by the costs, itself began to simplify. And from 2026, artificial intelligence: the European regulation enters into application in stages, and the firms' websites already offer models of algorithmic governance, ethics committees, registers of systems. Five waves, one product.
The structure is always identical, and anyone who has read the first two parts recognises it with the naked eye: an obligation formulated vaguely; a sanction feared more than applied; a sellable document that promises to shield from the sanction; no public measurement of what the document produces. On this structure the seller never gets it wrong, because he sells fear and he sells paper, and neither of the two has any quality control. The same firms, often the same professionals, garrison all five waves: the practice page lists 231, privacy, whistleblowing, sustainability, artificial intelligence, in a row, like vintages of the same wine. With an irony that deserves to be savoured to the very end: the consultants of the 231 model today sell compliance for the technology that is wiping out the value of their main product. The party does not end: it moves from room to room, and the orchestra is the same.
One wave alone contains, in embryo, the right lesson, and it is that of whistleblowing. There the legislator — European, not Italian — wrote verifiable obligations: channels with precise requirements, deadlines counted in days for replying to the person reporting, a public authority that supervises and can sanction. Much can be objected to in the implementation; but it is proof that writing measurable duties is possible even in this field, when the will is there. Decree 231, after twenty-five years, does not have a single obligation written that way. The comparison between the two legislative techniques — the measurable duty with the authority that checks, against the suitable model that nobody knows how to define — is the whole distance that separates a rule from a rent.
The two possible endings
From here, two scenarios, and it is worth looking at them without romanticism. The first is the end of the rent: the price of paper collapses, the market thins out, and value migrates towards the only thing that cannot be produced in an hour — real implementation. Checks in the field instead of matrices filled in in the meeting room; analysis of the actual incentives, the budgets, the internal rewards and punishments; serious internal investigations when the signal arrives; a Supervisory Body that costs money because it works; and, at trial, the ability to prove with data that the organisation was working. In this scenario the good professionals — they exist, and they know it — earn more, not less: they sell judgment, not photocopies. The second scenario is synthetic compliance: free paper does not kill the paper market, it floods it. Eight-hundred-page models generated in an afternoon, personalised in appearance down to the last comma, updated in real time at every new predicate offence; certifications on top of certifications; a judge submerged in documentation formally perfect and substantively empty, even less able to tell the difference than he is today. Cosmetics at zero cost, that is, universal cosmetics.
Which of the two endings comes to pass will not be decided by the technology. It will be decided, as always, by the referee and the rules. If the judge goes on reading the paper — and the paper will be perfect — the synthetic will win: hindsight will be exercised on impeccable documents, and will acquit or convict at random, as today, only with more pages. If the judge — instructed, equipped, trained — demands proof of implementation: registers of checks actually carried out, data on incentives, evidence of reports handled, uncomfortable decisions documented, then the paper will die altogether, and with it the party. The same fork applies to the legislator: a reward system that is content with the existence of the model chooses the synthetic scenario; a reward system conditional on measurable evidence of effectiveness chooses the other. The reform on the minister's table, as it is written, does not choose: and not choosing, in a market that is automating itself, is choosing the synthetic.
The buyer's questions
In the meantime, the fastest regulator remains the buyer, and to the buyer this part delivers a questionnaire. First question, to the supplier of the model: how many hours of actual work does this fee contain, whose hours, and how many of those hours have been replaced by an automatic generator — and to what extent does the discount reflect it. Second: which verification activities in the field are included — interviews, site visits, tests on processes — and which are instead documentary compilation. Third: on which real company data is the risk matrix built, and what distinguishes this model from the one sold to the competitor. Fourth, to the incoming Supervisory Body: what agenda, what budget, how many inspection visits a year, and what appears to have been done — not planned: done — in the previous year. Fifth, the most uncomfortable: what happens, contractually, if at trial the model is declared unsuitable — who answers for it, and for how much.
These are questions that no rule imposes and that any board of directors could ask tomorrow morning; the fact that almost nobody asks them is the exact measure of how far compliance is bought to be displayed, not to work. Because the paradox of the 231 buyer is this: he pays to protect himself from a finding of unsuitability, and never asks the seller to guarantee suitability — we are not aware of a single supplier of models who answers contractually for his own product. In any other market, a seller who guarantees nothing and a buyer who demands nothing would have less kindly names. Here they are called best practice.
What a serious system would demand
It is not difficult to say what would be needed; it is difficult to find anyone with an interest in asking for it. A serious system would demand, first, full publication of decisions: every 231 judgment and every dismissal in a public, free database, as any mature democracy does with its own precedents. Second, statistics: an annual ministerial survey of proceedings, outcomes, penalties, models assessed — the register of sanctions already exists, it need only be made to speak. Third, penalties and time limits brought to the level the OECD has been demanding for years: as long as an entity's offence costs less than an advertising campaign and is extinguished before reaching judgment, all talk of prevention is rhetoric. Fourth, real standards: public parameters of suitability, with rebuttable presumptions, but with certifiers subject to public audit and to revocation — the standard without control of the controller is the stamp on the photocopy. Fifth, training: organisational science in the education and continuing training of judges and lawyers, because one cannot judge what one cannot read. Sixth, transparency of conflicts: let those who write about Decree 231 declare their 231 engagements, as in any adult science. Seventh, and decisive: no rewards without measurement — whoever asks for discounts in exchange for organisation must accept that the organisation be measured, with indicators, independent verifications, counter-samples. Seven reasonable demands. Count how many of them the reform under discussion contains: none. And one innovation — the omission from the certificate — goes in exactly the opposite direction.
The test
This, then, is the test of seriousness, and it applies to every actor on the stage. For the legislator: a reform of Decree 231 that contains neither data nor transparency obligations is not a reform, it is maintenance of the rent. For the judiciary: to go on judging the organisation from the paper, after the paper has become free, means handing the trial over to the most verbose document generator. For the profession: the fee scales will survive only where there is something a machine cannot sell — and that something exists, but it is more laborious than the photocopy. For the academy: the twenty-fifth anniversary is being celebrated in the conferences, the admission of failure is on file at the ministry, and the window for becoming a science again is closing. The question in the title therefore has an articulated answer: the party will not end out of repentance, because none of the guests has an interest in switching off the music; it can end through obsolescence, because technology has reduced the value of the main course to zero; and it can end through external pressure, because the OECD and foreign regulators have begun to look inside the hall. The most likely outcome, in the short term, is a hybrid: the party will go on, to synthetic music, until someone — a judge, a legislator, a buyer — demands to see the accounts.
The final question
There remains the question that no reform and no machine can close, and it is the most serious of all: how could an entire legal culture — academy, judiciary, bar, publishing — dance for twenty-five years without ever asking itself whether the music existed? The answer does not lie in the law: it lies in the apparatus that has served as the law's garnish. The relentless conferences that never ask the forbidden question; the scientific journals devoted to a single decree; the prizes awarded on self-nomination; the chairs that teach the dogmatics of the model and not the crime the model is supposed to prevent. And, under the tree, the void: the criminology of the corporation, which in Italy does not exist; the harms — the deaths at work, the poisoned territories, the billions evaded — that nobody ever adds to the bill for the party; and a dogma that stands above everything: do not frighten the markets. With this — with the coloured, luminescent baubles on the Christmas tree, and the elephant the tree conceals — the final part is concerned.
A note on method
As declared in the text, this series was also prepared with artificial intelligence tools; every piece of data has been checked against an open source and the points that could not be verified have been indicated as such. The assessments of the reform concern the final report of the technical committee made public in early 2026: the political and parliamentary process may change it, and we will report on that. As with every whitecollarcrimes publication, any report of errors or inaccuracies will be verified and, if founded, accepted and corrected with due account given.
Main sources for this part
Technical committee for the reform of d.lgs. 231/2001 (d.m. 7.2.2024, coordinated by G. Fidelbo): final report and draft articles, made public in early 2026 (sistemapenale.it; giurisprudenzapenale.com) · Assolombarda, summary of the work of the reform committee (assolombarda.it) · Confindustria, position paper on the reform of the liability of entities, 11.4.2025 (confindustria.it) · Assonime, position papers 2025-2026 (assonime.it) · Bill A.C. 2632 (2025) on the Supervisory Body · Cass. pen., Sez. IV, n. 30039/2025 (certified standards and rebuttable presumption of adequacy) · Cass. pen., Sez. VI, n. 23401/2022 (Impregilo) · Trib. Milano, n. 10748/2021 (Monte dei Paschi; the Supervisory Body "an inert onlooker") · App. Venezia, n. 3348/2023 (Popolare di Vicenza; "osmosis") · Trib. Milano, n. 1070/2024 (acquittal for suitability of the model) · Osservatorio 231 at the Court of Milan (data on proceedings) · OECD Working Group on Bribery, Italy Phase 4 Report (2022) and follow-up · ISO Survey and Accredia data on ISO 37001 certifications · U.S. Department of Justice, Evaluation of Corporate Compliance Programs (justice.gov) · UK Serious Fraud Office, deferred prosecution agreements and published judicial approvals (sfo.gov.uk; judiciary.uk) · Loi n. 2016-1691 "Sapin II" and the Agence française anticorruption (agence-francaise-anticorruption.gouv.fr) · German draft Verbandssanktionengesetz (2020, lapsed) · Dir. (UE) 2019/1937 and d.lgs. 24/2023 (whistleblowing; supervision by ANAC) · Dir. (UE) 2022/2464 (CSRD) and d.lgs. 125/2024 · Reg. (UE) 2024/1689 (AI Act) · Observatories and professional press on artificial intelligence in law firms (2024-2026) · Commercial offerings of 231 prompts and assistants on subscription (survey, June-July 2026) · The previous parts of this series, with their respective sources.
whitecollarcrimes
a collective byline · guarantor: Avv. Luca Santa Maria · 11 July 2026