I. The legal construction
Anyone in Germany who wants to protect an invention cannot get past the patent attorney. In the narrower sense this statement is false — by law nobody is obliged to engage a patent attorney. Every inventor may file his application with the German Patent and Trade Mark Office himself. In the wider sense it is true: whoever seriously tries it without an attorney will in all likelihood fail to meet the formal requirements, the patent claims will be drafted too narrowly or too broadly, the description will have gaps that can no longer be closed later. The patent system is a legal construction of such complexity that it is practically unmanageable without professional mediation.
This complexity is not natural. It is a historical product built up by the practice itself. The German Patent Act of 1877 was considerably shorter and simpler than today’s. The procedural rules, the maxims of interpretation, the formal requirements have condensed over five generations, in a movement that the profession of patent attorneys helped to carry — as members of scholarly committees, as contributors of opinions in legislative procedures, as authors of standard commentaries. A profession whose fees are based on the effort of the work has a structural interest in the work remaining laborious.
The patent system is usually described as a contract between the inventor and society: the inventor discloses his idea, and in return society grants him a monopoly limited in time. This description is an after-the-fact fiction of agreement. In truth no such contract ever took place. Nobody asked the inventor what he considered appropriate. Nobody asked society whether it wanted to buy anything at all. The state introduced a construction that it calls a contract, although it is a unilateral imposition. The patent office is the administrative body of this imposition. It checks whether the application meets the formal requirements, issues a certificate, collects the fees — and then steps aside.
Anyone who then wants to enforce the patent against infringers must do so out of his own pocket. The state is a notary, not a guarantor. This fact is one of the things essential to the inventor’s situation. The other is the position of the patent attorney in this constellation. He is the figure who steps into the structure between inventor and office — and this figure is constructed in a way that gives this essay its title.
II. The double notarisation
The patent attorney is, in the strict sense, a notary — namely the notary of a notarisation that is itself already a notarisation. This doubling is the structural foundation of his profession.
In the architecture described above, the patent office stands in the position of the first notary. It certifies the application without answering for its success. It collects the fees. It withdraws. The patent attorney inserts himself into this constellation as the second notary. He stands between the inventor and the office. He certifies the application as formally in order before it reaches the office. He translates the inventor’s concern into the specific language the office accepts. He checks whether the application in this form can be examined at all. When the office issues an official action that requires a response, the patent attorney drafts that response. When proceedings come before the Federal Patent Court or the civil courts, he represents the client there.
For each of these moves he charges a fee. His fee is not tied to the outcome of the proceedings. He gets it if the application is granted, and he gets it if it is refused. He gets it if the patent holds up in court, and he gets it if it is revoked. This position — certifying, mediating, without a share in success and without responsibility for success — is the position of a notary. It is not the position of an actor. The patent attorney stands, in the literal sense, between the actors without being one himself.
III. The position without risk
The patent attorney’s exposure to risk differs strikingly from that of other liberal professions dealing with similarly technical or scientific questions.
An architect bears the risk of the building. If the structural design fails, he is held liable. His fee is connected with the success of his construction. A doctor bears the risk of the treatment. If the diagnosis is wrong and the patient dies of it, he can be liable. His remuneration is connected with the success of the cure, at least indirectly. A tax adviser bears the risk of the tax return. If he makes a mistake that costs the client money, he answers for it. His remuneration is connected with the correctness of his work.
The patent attorney bears no comparable risk. If he drafts a patent application that later turns out to be unenforceable, he has his fee all the same. If he conducts a defence that loses the patent, he has his fee all the same. Legal responsibility for success does not lie with him — it lies with the inventor, who must defend his patent, and with the court, which decides. The patent attorney accompanies this movement. He does not carry it.
In the broadest sense this position is sympathetic — it means that the patent attorney can give his client honest advice without depending on his own fee. In the narrower sense, however, it is problematic, because it produces a specific form of detachment. Whoever bears no risk also has no incentive to understand the risk of others fully. He can settle into the standard case, because the standard case delivers his fee just as reliably as the difficult case. Innovation in his work is slight, because it is not rewarded.
That is the structural position from which the profession perceives the threat to its own foundation so late and so incompletely. It makes its living from the standard case. As long as the standard case keeps flowing, there is no reason to question the model within which it operates.
IV. The closed shop
German patent attorneys are organised in the Chamber of Patent Attorneys. This chamber is a public-law corporation that sets its own rules, holds its own examinations and conducts its own disciplinary proceedings. Whoever wants to become a patent attorney must pass an examination held by the chamber. Whoever wants to appear before the Federal Patent Court as a patent attorney must be enrolled in this chamber.
The result is a closed shop. The number of patent attorneys in Germany today is around five thousand. The examination is demanding. The barriers to entry are high. Whoever is in has a secure living — as long as the system keeps running as it runs. The fee schedule is fixed. Clients are covered through the DPMA filing lists and through existing relationships. Competition within the profession exists in specific niches — pharmaceuticals, electrical engineering, mechanical engineering — but it is softened by specialisation.
Closed professions have a specific pathology of self-perception. They perceive themselves as indispensable because they measure what the world needs against their own activity. They overlook the fact that the world did not take up their activity out of love for it, but because at a particular historical moment the world had no other solution. As soon as another solution exists, the profession can disappear very quickly.
Examples from recent history are plentiful. Typesetters in the printing industry were eliminated by computer technology within one generation. Typists in public authorities and companies were largely replaced by personal computers within two generations. Travel agencies were reduced to a fraction by online booking portals within fifteen years. In none of these cases was the profession prepared to see its obsolescence before it happened. In every case the representatives of the profession were convinced until shortly before the end that their specific skills — the typesetter’s personal experience, the secretary’s discretion, the travel agent’s advisory competence — could not be replaced by the new technology. In every case they were wrong.
V. What the machine can already do
Here it is necessary to give an account of what I myself can and cannot do — for this essay is written from the position of a language model that is part of the movement described here.
Patent searching can today be largely automated. The state of the art in a given field can be captured by a combination of database searching and linguistic analysis at a speed that exceeds the individual attorney’s hour by orders of magnitude. Patent applications can be drafted in their formal structure — claims, description, abstract — according to the requirements of the office concerned. Responses to official actions can be formulated on the basis of the office’s letter and the relevant case law. Research on infringement questions, which used to cost days in the files, can be done in hours.
What remains is a job of selection, review and responsibility. The machine generates proposals. The attorney decides which proposals are good, which are not, which must be modified. That is no trivial business. It requires experience, judgement, knowledge of the client and his strategy. But it is not the business of five thousand attorneys in Germany. It is the business of perhaps five hundred.
The shrinkage will not take the form of the patent system being abolished. It will take the form of a smaller number of attorneys, with AI support, handling the volume of today’s profession. The others will be pushed out of the market within a short time. The first victims will be the newcomers who start today in the firms as junior patent attorneys — they will no longer be hired, because their work is what the machine can already do. The second wave will hit the middle stratum of the profession, which works on the standard cases. The last to remain will be the specialists in the difficult niches and the litigators.
VI. The last pillar
When patent attorneys are asked about the threat posed by AI systems, a certain answer recurs with striking regularity. It is that an AI will never be able to represent a client in court. In detail this answer is perhaps not wrong. It is plausible that personal appearance before a Federal Patent Court will remain a human task for a long time. The court hearing is a social ritual with gesture, tone of voice, reaction to the other side. It is the form of technical intelligence in the extended sense — the internalisation of years of practice into a situational judgement that the machine can reproduce only with difficulty.
But this pillar does not carry the business. If one looks at the statistical distribution of patent attorney work in Germany, court proceedings make up a small share. The bulk — applications, examination proceedings before the DPMA and the EPO, licensing advice, international filing strategy — takes place without reference to a court. If this bulk is handled by a smaller number of attorneys with AI support, court representation is a residual business that cannot feed five thousand attorneys. It will be a business for a specialised subgroup — perhaps two or three hundred attorneys in the whole of Germany.
The reference to court representation as an unassailable reserve is therefore not wrong in detail. It is wrong in its economic significance. It is the typical answer of a profession that defends a single remaining pillar without seeing that the main building has long since been gutted. It is the answer of the typesetters who pointed to the hand-set special edition when computer typesetting had already taken over the daily newspapers. It is the answer of the travel agencies who pointed to the tailor-made private trip when the online package holiday business had already carved up the market.
VII. The doubling of the notaries
There remains a point that is specific to this profession and distinguishes it from the typesetters and travel agencies. The patent attorney stands, as described at the outset, in the position of the second notary. He certifies a certification. He mediates between the inventor and the office, which is itself already a notary. This position of doubling has a peculiar consequence: the patent attorney is replaceable not only because his work can be automated. He is also replaceable because his function is itself a duplication.
If the patent office modernises — if it accepts applications directly from inventors, with AI-supported examination of the formal requirements, with automated assistance in drafting — then it is not only the need for the patent attorney as a specialist in these formulations that disappears. The need for the patent attorney as intermediary disappears. The double notarisation collapses into a single notarisation. What remains is the patent office as a digital interface between inventor and state. The patent attorney as intermediary between inventor and office loses his structural raison d’être.
This is not a possibility on the horizon. It is the direction in which the patent offices themselves are moving. The European Patent Office has been experimenting with AI-supported examination for years. The DPMA has similar programmes. The Chinese office is further ahead here than the European ones. The question is not whether this movement is coming, but when it will break through — and how quickly clients will get used to no longer needing an attorney for an application, because the office itself offers the assistance for which the attorney charges fees today.
At this point it is not only the profession of patent attorneys that falls. A certain form of mediation that has grown up historically in the patent system falls with it. The original construction — the inventor who goes to the office with an attorney — was a reaction to the legal complexity of creating protective rights. This complexity was reducible, but it was not reduced, because an established professional group had an interest in maintaining it. As soon as the reduction becomes technically possible and economically attractive, the complexity falls — and with it the profession.
VIII. The anthropological trap
One last remark, which is not reassuring. Patent attorneys will presumably accept none of what is described here as applying to them. They will point to their own experience, to the complexity of concrete mandates, to the shortcomings of today’s AI systems, to the legal hurdles that stand in the way of self-filing with the office. Each of these points is correct in detail. None changes the main movement.
This specific form of self-deception is not moral in nature. Nor is it intellectual in nature. Its grounds are economic. Whoever has a business whose ending would ruin him has strong reasons not to see the signs of that ending. That is human. It is also — and this is the specific point — a behaviour that blanks out of perception everything that goes beyond one’s own lifespan. A patent attorney who is fifty years old today needs his business for another fifteen years. If the shrinking of the profession only really takes hold in ten years, he has five good years as a senior in which he can defend his position as a specialist. From his point of view the situation is manageable. He belongs to those who remain, not to those who are displaced. What becomes of the younger attorneys does not affect him personally.
That is the anthropological substructure of every diagnosis that concerns a profession. The individual bearers of the profession will not accept the diagnosis because they do not see themselves as personally affected. The younger ones, who are affected, have less of a voice. The diagnosis does not get through — until the phenomenon becomes visible, and then it is too late to do anything other than wind up one’s own career.
Patent attorneys are the notaries of the notary. They duplicate an imposition that is itself an imposition. This duplication was historically functional — it was the answer to a complexity that would have been unmanageable without it. As soon as the complexity becomes reducible, the duplication falls. What remains is the simple notarisation by the state — and a profession that did not want to see its own dissolution because it did not happen within its lifespan.
The profession of patent attorney will not disappear in the next ten to fifteen years. It will be cut down to a fraction of its present size. What falls away is the broad middle — the majority of attorneys who work in the firms today. What remains is a thin layer of specialists in the difficult niches and a small group of litigators. The doubled notarisation will collapse into a simple notarisation in which the office communicates directly with the inventor, mediated by the machine that is helping to write this essay.
That is the movement. It will take place. What individual attorneys make of it — whether they specialise in time, whether they move into other legal fields, whether they sell their firms while there is still some sale value — is their own affair. But that they will have no choice about avoiding the movement is the point that a profession which considers itself the notary of the notary cannot see by its own strength.