Machines, artificial intelligence and IP departments

10-10-2017

Volker Spitz

Machines, artificial intelligence and IP departments

jirsak/ iStockphoto.com

While artificial intelligence is unlikely to replace IP departments overnight, it may have an increasingly influential role in the medium and long term, as Volker Spitz of Brandstock finds out.

The ability of machines to perform tasks better than humans is not new. A simple calculator is able to process numbers far faster than most people, and with a 100% accuracy rate. It is this combination of speed and accuracy that makes it inevitable that many jobs which currently require human input will in time be done by machines. And as artificial intelligence (AI) becomes more effective, even things which we assume require human input may also in time be done by machines.

Of course, this has implications for IP departments. Taking a snapshot of a typical IP department at the moment, we can see that already, many of the tasks that used to be done by humans are being undertaken by machines. In this article, we will attempt to map the functions of an IP department and consider what is already being done by machines, and where AI is likely to have a role in the future.

Maintaining IP rights

One of the key responsibilities of any IP department is to understand the IP assets that the business holds and organise them effectively, keeping records updated and maintained in order to ensure there are no weak spots in the portfolio. In the not too distant past, that would have meant paper filing systems and an extensive paper diary, coupled with a bill for postage that would make most modern IP departments shudder.

Today, machines can shoulder much of this burden. In-house docketing systems remove the need for extensive paper filing, and keep track of the various stages of applications, renewals, recordals and so on.

However, while this advance in technology has rendered some of the work that used to be done by humans obsolete, it has also increased the ability of IP departments to work efficiently and probably created more useful new work in the aggregate.

Currently, however efficient these processes are, they still rely on human judgement to underpin them. No machine is currently able to assess what are the key assets in a given portfolio, or where are the most challenging jurisdictions, for example. Humans need to make these decisions.

That is likely to be the case for a few years yet, but it is possible that in the future, even some of these assessments can be taken over by artificially intelligent machines. It’s possible, in theory at least, for an AI system to assess (and indeed, learn) which jurisdictions are likely to be key for a given asset. An internet connection and a set of parameters on which to base that judgement could achieve that, assuming that the relevant information on the market, the patent office and the business is available and readable by a machine. This is unlikely to be possible soon, but it will probably be feasible eventually.

Advice

Another key responsibility of an IP department is to advise a company on what IP rights are necessary, what the risks are, and where in the world they are needed. This is currently a very human task, and is likely to remain so at least in the medium term. Advice requires a nuanced understanding of the needs of a given business, including the personalities and goals of the senior management within it.

While it’s hypothetically possible that an AI machine could perform some of these tasks, especially with regard to the technical aspects of filing applications, it would require a level of intelligence and learning well beyond what’s likely to be possible for the foreseeable future in order to completely supplant the human element.

However, machines can already help you at the application stage, for example with searches. AI will almost certainly play more of a role in that as time goes on, and indeed Brandstock is already exploring how to integrate AI into their search services.

Cost control

One of the key reasons for moving certain functions to machines is the potential cost-savings available. These savings could include lower staffing costs, but the savings also extend to the costs of managing assets. For example, a machine solution that improves a business’s effectiveness at renewing its IP assets, or which ensures that application deadlines are met, is sure to save not just official fees but the costs of employing agents to rectify issues in particular jurisdictions.

Brandstock uses sophisticated software solutions to help it manage the application, renewals and recordals pipelines to provide certainty that costs are kept to a minimum at the same time as risks are removed from the process.

IP is an unusual asset: in many companies it becomes big news only when there’s a failure. If you have trademarks registered and up to date in the jurisdictions you need, business can continue. But if you find a trademark is not filed or up to date when you actually need to enforce it, then you have a big problem. Machines already can remove some of the burden from humans, and AI should enable these processes to be even more effective in time, since a system that could be taught and act on information about patent office processes in different jurisdictions, for example, would be one that could remove human error from the process. Such a system would reduce cost by removing humans from the chain, but would also reduce it by de-risking the process.

Licences

IP departments often take responsibility for negotiating and drafting licensing agreements, and managing the transfer of IP assets into and out of the business as required.

This used to be a potentially inexact science, relying on individuals’ assessments of the value of a given asset, its potential performance in the market, and its longevity as a licensing proposition. With the right market conditions and multiple potential buyers, this approach works well, since in theory the market is fairly effective at dictating prices, and the nuts and bolts of an actual licensing agreement can be fairly standardised.

However, especially in smaller or more niche markets, this is a risky way of managing licensing. Machines and big data can help enormously in negotiating and drafting agreements, especially as pricing becomes more transparent. It is already possible to gain a more accurate valuation of a particular IP asset by using software to analyse comparable assets in a market, look at historic licensing agreements and establish the economic context for the asset.

There are already significant databases covering licence agreements, royalty rates and valuations. As these grow in accuracy and size, they should become exponentially more accurate in enabling you to assess the value of an asset, since each new licensing deal contributes more data to the aggregate and improves its effectiveness.

Similarly, machines have an important role to play in managing and enforcing royalty agreements, especially when partnered with an overall system for keeping the assets up to date. In the future, it’s possible that an AI system could identify where licensing is desirable, identify appropriate terms, and even adapt standard agreements to generate contracts for them. It will almost certainly be a long time until humans are entirely removed from this process, but the labour- and time-intensive parts of it can already be managed to some degree by machines, and that trend will undoubtedly intensify.

Maximising the value of a portfolio

Linked to licensing is the role of an IP department to ensure that a business makes the most of its IP assets. Many companies own patents from which they derive little or no benefit—legacy filings that are neither being worked nor licensed. Unless they serve a defensive purpose, such assets are a pure cost to the business. Trademarks too may have value beyond what they are being used for, perhaps through potential licensing in new jurisdictions or for new purposes.

Using machines to understand what’s in your portfolio, what is generating revenue and what is essentially useless, is an effective way of identifying assets that might potentially be licensed or sold.

The success of this approach ultimately rests on the accuracy of your understanding of your own assets. Once you understand what you have, you can make sure it’s all up to date. With that information embedded into the business, you should be in a better position to identify opportunities for future revenue streams.

AI is not going to replace IP professionals overnight. In the medium term, it is likely to enhance human capabilities rather than supplant them. But as the machines become able to take on more and more tasks, and to do them both more quickly and more accurately than their human masters, it is sure to make for an interesting future in the profession.

Volker Spitz, Brandstock, artificial intelligence, IP departments, AI, licensing, AI system, machines, portfolio, cost control

WIPR