The Rise of AI Patents

Artificial Intelligence (AI) is no longer just hype; it’s the reality for thousands of businesses and millions of consumers. At least, that’s what we conclude when we see stories like this one from Enterprise Tech that dives into the data of patents. According to the article, IFI looked into AI patents and “estimates the patent category that includes machine learning and neural networks grew at a 34 percent annual rate over the last five years” making it the fastest growing category in the patent database, and pattern isn’t slowing down either.As you’d expect most of the patents come from large companies like Google, IBM, Facebook, and Microsoft, but that number includes those from companies like us here at Attivio where we are constantly updating our own patent portfolio. Taking a deeper look around the 2017 patent numbers, PhD student Alexej Gossmann found that when he searched the core patent database for keywords, including “machine learning”, “deep learning”, “neural network”, “artificial intelligence”, “statistical learning”, “data mining”, or “predictive model”  within the titles, descriptions, or claims text, he found 6,665 patents that met his criteria. That’s of a total of more than 320,000 utility claims made last year.Of those, the vast majority fall into the category of “neural network” or “machine learning.” In fact, IFI noted a similar trend, finding that the large companies seem to be focusing specifically on machine learning designed to train predictive models. Attivio's AI PatentsAs for our own patents, Attivio holds ten that As is the case with any patent, it often lags behind the innovation itself. Gossmann found that the average patent takes about 2.5 years to move through the system, though some move much faster than that. While we would love to believe that every AI-related patent is of the highest quality, like anything else, the world is filled with those of questionable value.  With AI out in the lead, what’s next? IFI says the next fastest growing area is automated vehicles, which happens to use much of the same AI-based technology.