Report: AI Research Slowing Down, Development Increasing

Advertisement
Advertisement

Research in the artificial intelligence (AI) segment has been slowing down in recent decades, with major tech giants primarily focusing on developing existing solutions and technologies, data compiled by the Golden Goose Project suggests. While research and development (R&D) budgets are generally approved together, the latter is reportedly getting much more attention from tech executives these days. Research is still moving forward, albeit at a significantly slower pace, with tech companies now being more willing to rely on underfunded universities and startups to handle advancements in the AI field.

Duke University's Fuqua School of Business illustrates those recent trends through the aforementioned Golden Goose Project that attempts to visualize recent priority shifts in the industry, though it's primarily based on patent records and research achievements statistics that may not be a perfectly accurate technique of measuring such trends on their own. The project still strives to quantify the initially compiled data and determine how AI research is applied to development, looking at both internal company implementations and commercialization initiatives. The graphs that can be seen beneath this writing and the project's other findings hint that startups are becoming an increasingly more important factor in the process of commercializing research and bringing it under the umbrellas of corporations, primarily through acquisitions. As a consequence of this state of affairs, corporate giants are able to maintain their innovation achievements while doing significantly less research and redirecting those funds to development, the data suggests.

It's still debatable whether the current scenario is sustainable in the long term, especially as universities are still strapped for funds and most AI research endeavors are government-funded, thus often highly politicized, some industry watchers believe. Private funds can alleviate the situation to a degree but are still unlikely to completely neutralize the problem, at least for the foreseeable future. Apart from the aforementioned method-related shortcomings, the Golden Goose Project may also not be perfectly representative of reality due to the fact that its visualizations are based on data from between 1980 and 2006, and with the likes of Google's DeepMind and some other private corporations now heavily investing into the AI segment, the current situation may not be so grim. AI as a whole has been in development for decades now but is still in its infancy, at least in the context of the technology's potential applications. Current commercialization efforts mostly come down to digital assistants accessible through smartphones and Internet of Things (IoT) devices, though it remains to be seen where the industry goes on from here.

Advertisement