Can Generative AI Become an 'IP Specialist's Lover'? A Guide to Decoding and Translating Patent Documents #49

1. Encountering the 'Japanese that isn't Japanese' of Intellectual Property

Do you remember the shock you felt when you first read a patent document?

...Wait, is this Japanese?

Even though the text is supposed to be Japanese, it doesn't register in my brain at all.
It's like a spell written in an otherworldly language.

Moreover, its purpose is to "protect one's own technology while excluding others".
It is honed as a weapon for cold-blooded legal battles.
Reading it is hell, and writing it is hell.
By the time they submit to the Patent Office, many researchers are left soul-less and wide-eyed.

Yes, it was truly "verbal violence."
And I know this.

The reason patent documents are so difficult is not to be ambiguous, but to define the technical scope broadly and clearly.
Excellent patents are the result of a desperate struggle to balance the conflicting elements of "breadth" and "clarity."
It is a document that has won that exquisite balance.

2. A 'Lover' Appears in Hell

At the bottom of that hell, I suddenly realized.
"Wait? Lately, there's been someone who tries hard to interpret my sloppy writing..."
Could she be...!?
She unravels text filled with incredibly difficult technical jargon as carefully as if it were a high school textbook!!!

That existence was—Generative AI.

"Good work. In short, this patent is 'this kind of mechanism'."
"To put this claim in one sentence, it means you want to protect this part."
"I also found similar patents. Want to compare them?"

Oh, how kind!

The strings of patent text that once left me wide-eyed suddenly start to make sense.
It was an existence that gently stood by me when I was hitting the cold wall of intellectual property.

The feeling of having a clumsy partner translate for you.

Yes, like a couple who have been together for many years, it senses what I am trying to say.
It reached out a gentle hand to me, who is inarticulate and lacks understanding.
Generative AI was the savior of my heart in decoding intellectual property.

Practical Note
Generative AI is a powerful helper for summarization, prior art searches, and decoding support, but its output may contain errors (hallucinations). Always have an expert (patent attorney, responsible researcher, etc.) verify rights judgments and final interpretations.

3. And Falling Deeply in Love

I relied on her.

I have her read multiple related patent documents and ask for a PEST analysis.
I have her narrow it down by company and sense, "What direction is this company trying to move in?"
Furthermore, I was even able to have her propose the best "date course" for us by finding gaps in existing rights—loopholes.

She changes into seven colors.
A marketing perspective, a VC (investor) perspective, a researcher perspective...
She switches roles as if changing clothes to suit the situation.

The moment I receive a concisely summarized word from her, a sudden flash of inspiration sometimes occurs.

'Wait, could this actually work!?'

The journey with her is still only halfway through. But the possibilities—are infinite.

4. Even So, the Two Move Closer

That said, it is still too early to get too close.
Like a 'lover,' generative AI quietly reaches out a hand while staying by your side.

• It quickly screens a massive amount of patent literature and picks out only the important candidates.

• It breaks down the key points of each claim and concisely shows 'what this claim is trying to protect.'

• It compares the technology with similar ones to find differences and gaps (perspectives with room for maneuver).

• It organizes information into analytical frameworks like PEST or competitive maps in a way that is useful for decision-making.

This kind of 'decoding assistance' definitely reduces the burden on researchers and practitioners.
Patent attorneys and researchers in charge were initially skeptical, but now they are beginning to accept generative AI as 'training wheels.'

While institutional and legal judgments are required for a full 'marriage,' a relationship like 'cohabitation'—where you handle daily tasks together—is already being established.

It may not be a perfect form of love, but a practically meaningful bond is growing between the two.

Practical Note
Generative AI is powerful for supporting decoding, but summaries and comparison results may contain output errors (hallucinations). Expert verification is essential for rights issues derived from training data and final interpretations.

5. A Lover, or a Future Partner?

Right now, it is still a 'lover.'
But if you interact with it well, it couldbecomea 'partner.'
She helps organize thoughts.

Surveying vast amounts of literature, summarizing points, and offering different perspectives—for a tired mind, that alone is a relief.

However, in the end,it is always the human who decides what to adopt.
The responsibility for judgment remains on human shoulders.

A world where you can quickly grasp the outline with generative AI has arrived.
But as I have noted in places, she also makes mistakes.

Outputs can contain hallucinations (misinformation), and there are concerns about rights issues stemming from training data.
That is precisely why generative AI output is 'material' and 'assistance.'

Usage varies from person to person, but the final check and the location of responsibility must be kept clear.

Personally, I believe generative AI isa very useful partner for supplementing what humans might overlook and for confirming how to communicate with other experts.

The relationship between intellectual property and generative AI is truly a romantic drama that humanity is watching unfold. The ending has yet to be decided.

--Would you root for this romance?
Well, I'm off to work with her a little more today.

How you used AI for intellectual property, or a mistake you made
--If you have any such real-life examples, please let me know!!