The weblog of the Journal of Intellectual Property Law and Practice. Here's where editorial panellists, readers and contributors can come together and share their view on all aspects of IP law and practice. Join us!
What future for robots and AI in trade mark practice? Here's our February Guest Editorial
How can Artificial Intelligence help in the trade mark practice field? JIPLP Editorial Board member Darren Meale pens this month's editorial.
love trade marks. I live and breathe them. I love helping people understand
them, get them and fight for them. I’ve learnt the old, established, sensible
ways of doing this. But I also don’t like standing still. No one else is – the
world is changing. Here is my brief vision for the future of trade marks.
The EU Intellectual Property Office (EUIPO)
issued nearly 7,000 opposition and cancellation decisions in the past 12
months. How many did you read? Can you tell me how many times in the past five
years word marks differing by only two letters were found to be confusingly
similar, and how many times they were found not to be? The EUIPO’s similarity
tool records 118,073 decisions on the similarity of particular goods and
services with other goods and services from 19 trade mark offices. How many can
Trade mark lawyers draw on years of
experience and judgment formed from that experience to make wise decisions on
the suitability of marks for registration; on clearance searching; and on trade
mark conflict. They will confess, or at least purport, to be au fait on
the local trade mark office’s latest practice, and their attitude towards
findings of likelihood of confusion. In which direction has the pendulum swung?
Has the registry adopted a pro-brand owner stance, or might things be turning
back towards a more liberal approach? We have a view.
Further, trade mark lawyers know the rules.
It has taken us a while to learn them, and learn how to apply them, but we know
to make a visual comparison, an aural and a conceptual one. We know to discount
the non-distinctive elements and take greater note of the distinctive ones. We
wrap these results up in a few moments and we apply our human instincts and we
make our judgment. Some of us no doubt believe – and perhaps rightly – that
because we have worked with genuine dedication and passion in trade marks for
many years that our instincts are pretty good, and we sell our legal services
accordingly. But how long will our superior judgment reign supreme?
A computer does not work like a human. It
cannot see the subtlety of language, it struggles with nuance, and it can only
operate in a mechanical, unthinking way. It cannot put itself into the mind of
a human being and try and work out whether he or she will be confused by PEPSI
v POPSI or GOOGLE v GIGGLE. It cannot prepare an overarching strategy. Well
maybe not today, so don’t panic. But what about tomorrow? The computer can
certainly be taught rules. Once it knows the rules, it will apply them in
exactly the same way every time. And it will do it at effectively the speed of
light. It does not need to read 7,000 EUIPO opposition and cancellation
decisions. It has just read them. If we teach it how to interpret those
decisions, it will do so just fine. I do not pretend to understand fully how AI
and machine learning work. But I do recognize the principle that the more data
we make available to a computer the better it gets at making judgments on human
behaviour. The same will inevitably go for making judgments on trade mark
If you have such an AI now – please get in
touch, as I’d love to see it. But one day we’ll get there. Your client will
cease to ask “what do you think?” and begin to ask “what does the computer
say?”. You, the lawyer, won’t be out of a job as such, but your job will be
different. Now you program the computer and you help it learn and you interpret
what it says. You add a human skin, because human relationships are essential
to good client/lawyer relationships. But you no longer rely on your judgment,
or your instincts based on the last five cases you read. You rely on the
computer, which has read all 63,000 of the cases in the last 10 years, has all
the data from the registries of the 117 countries signed up to the Paris
Convention, has spotted all the trends. You rely on the computer not because it
is faster or perhaps cheaper for the client. You rely on the computer because
it is smarter: it gets the answer right more often than you.