Dedicated AI chips have been a fixture on high-end smartphones and some mid-range devices for several years now, dating back to Huawei’s Mate 10 series offering a so-called neural processing unit (NPU) in 2017.
Since then, we’ve seen MediaTek, Qualcomm, Apple, and Samsung all adopt dedicated AI silicon on their smartphone chipsets. But do you think these AI chips have lived up to the hype though? That’s the topic of our featured poll today, so choose an option below.
We’ve seen smartphone companies tout AI silicon for improved scene/objection recognition, faster and higher quality image processing, security, and better system performance. So there have been a ton of claimed use-cases for the technology. And Google is pushing even more use-cases for the Pixel 6 series, like offline voice dictation/translation.
Then again, today’s smartphone chipsets are very adept in general even without dedicated machine learning hardware. Faster CPUs, GPUs, and ISPs combined with improved software and algorithms can enable some of these use-cases too.
So do you think AI chips have failed to meet manufacturers’ claims? Let us know via the poll above and leave a comment if you have more on your mind.