Inventions becoming inventors… It might sound like sci-fi, but it’s not really such a far-fetched idea. In fact, it’s a prospect that’s already becoming reality, given the recent news that an AI-created medicine for the treatment of OCD will be tested on humans for the first time.

While the world’s first patent applications for machine-designed inventions were rejected in January by the EPO, this development brings fresh attention to the AI-inventor debate.

What may seem to many like simply an interesting experiment might well have far-reaching implications. Could this be the tipping point when technology goes from being a facilitator and an enhancer of human endeavour, to a developer of innovation in its own right? If it is, the impact on creative individuals and businesses, and IP owners at large could be momentous. If there’s one thing we can say with certainty, it’s that it will pay to be prepared for this kind of brave new world.

Revolutionising the Creative process

Driving this move towards ‘robot’ invention is the combination of big data and AI. The aggregation of data points from across the global IP ecosystem – such as volume of patent filings for a particular product in a particular jurisdiction for instance – provides an intelligent map of potential risk and opportunity. Adding to this AI (or at least machine learning) which can quickly scan and derive insight from the data set, creates a hugely powerful combination with the potential to revolutionise the invention process.

We’re already seeing machine learning and AI being deployed to help humans identify gaps in the market and seize new commercial opportunities, for example by scanning vast pools of IP data for existing patents that could be combined in a new way, or identifying new opportunities to license in a specific market.

From this, and in light of recent news, it isn’t really such a stretch to imagine a machine anticipating new sources of demand for new products and services. And, in addition to creating new inventions, intelligent machines of the future could run tests to evaluate everything from initial concepts through to finished designs.

These are the kind of cutting-edge solutions that have the potential to be a major competitive differentiator and efficiency driver for companies.

That said, although sentiment may change in the future, for now, the idea of robots taking on what has always been considered to be a uniquely human trait – the ability to be artistically/intellectually creative – is a tough one for many people. Customers may need to come around to the concept if they are to buy into it. Machines may already be writing music, creating lyrics and melodies protected by copyright, but how receptive will listeners be? How will we feel if robots win Nobel prizes? Ultimately, the ends will need to justify the means.

re-routing the ip pathway

For maximum commercial benefits to be realised, current IP systems will need to adapt – and probably quite fast, given the pace of change we are seeing in the development of AI and machine learning today.

For instance, current IP laws generally (for instance in the US and UK) require creations to be assigned to a human. For example, copyright law requires a person to be an author or artist, and patent law requires a named human inventor for the process of idea protection to even get off the ground. The prospect of a robot creating and owning IP will require a sea change in how the regulatory and legal system is imagined. Decisions will need to be made as to who should speak on the AI inventor’s behalf when it comes to the management of IP rights, for example concerning the licensing of IP rights to third parties or whether patents should be renewed. And since human individuals and teams will often have worked with the AI creator, how will the software (i.e. AI) developer profit from AI-derived IP and to what extent does the robot, or the robots owner / developer, stand as liable should an invention malfunction or result in an accident? Just look at the debate raging around the issue of liability when it comes to self-driving vehicles as an example.These are complex and thorny issues and, as with much technological innovation, the regulatory landscape is seeking to play catch up. The USPTO is, for example, consulting on the ‘patentability’ of artificial intelligence inventions.

As IP ownership issues look set to become more complex, consideration needs to start being given to these kinds of questions now, both by IP rights holders and by legal advisers specialising in intellectual property.

AI and machine learning are already shaking up many industries – often for the better; streamlining the invention process, providing competitive intelligence and generating new sources of value. But their role as a driver of new ideas and creator of new products adds another dimension altogether and is likely to form a key component of the future of invention which, at CPA Global, we call New IP. Innovative companies large and small will be keeping a close eye on these developments to ensure that they are well placed to benefit from, and not be displaced by, the power of AI. The reality is that this is no longer fantasy. The technology is there, just waiting to be unleashed.

This article is part of CPA Global’s New IP campaign. To find our more, visit