Microsoft (MSFT - Free Report) recently voiced its concern over the ongoing use, misuse, abuse and overuse of facial recognition technology, via a blog post by Microsoft President, Brad Smith.
Facial recognition technology is increasingly being leveraged by various public and private companies, government agencies, among others, to fulfill their respective requirements. The use of the tech is almost infinite, for instance, unlocking screens, identifying and uniting family members, receiving suggestions for tagging a person on a picture, to mention a few.
As technology advances, the real and virtual slowly gets entwined and there is no way around it. Consequently, Microsoft calls for regulation of a few aspects of the technology in order to protect the very basic ideas of freedom and privacy. To start off, approval of the people coming under the scanner are often not considered. Microsoft remains highly concerned with this violation of a fundamental right, which has been going on for quite a while.
The next significant issue is the credibility of the technology. Facial recognition technology accompanies a considerable bias inthe way it functions. Notably, various studies have proven that the technology is surprisingly erroneous.
Biased AI: A Major Concern
The most recent breakthrough study is from MIT’s Joy Buolamwini which focused on facial recognition application, powered by Artificial Intelligence (“AI”). The companies in focus were IBM (IBM - Free Report) , Microsoft and Face++. Darker-skinned females formed the most unrecognized group, highlighting significant error rate of almost 34.7%. Per the study, “The maximum error rate for lighter-skinned males is 0.8%.” This significant disparity is a matter of concern.
The study has been acknowledged by the technology fraternity at large. Following the research, an increasing concern has emerged from the tech giants to reduce the biases, technically the errors. The tech companies are continuously working on APIs that remove or at least minimize this bias.
IBM had almost immediately addressed the issue by issuing a new API which recognized darker females with reduction in error to 3.5%. Meanwhile, Microsoft in the past month announced an updated and improved Face API available on Azure Cognitive Services, which reduced the error rates for both men and women having darker skin “by up to 20 times”.
What the Investors Need to Know?
Microsoft has returned 46.5% in the past year, substantially outperforming the industry’s rally of 35.8%. This outperformance can primarily be attributed to its continuous efforts in strengthening AI and Internet of Things ("IoT") based capabilities. Undisturbed focus on Azure continues to remain a key catalyst.
The improvements in sensors, cameras, and strengthened machine learning (“ML”) capabilities are paving the way for companies and government agencies to increasingly leverage facial recognition technology. Per MarketsandMarkets data, the global facial recognition market is envisioned to grow at a CAGR of 13.9% to reach$7.76 billion by 2022 from $3.37 billion in 2016.
However, Smith necessitates the need to go slow on the use of facial recognition technology in a bid to be socially responsible. The company aims to use facial recognition in a transparent manner, not impeding any individual’s fundamental right to privacy. Reducing bias component is expected to be a key catalyst, going forward.
Bias in AI has been doing the rounds for quite some time now. The companies, especially the tech behemoths are aiming to minimize errors relating to biased outcomes.
For instance, IBM in its “5-in-5” 2018 goals highlights that though bias in AI may explode, it is the unbiased AI which will survive the long run. After the flak Facebook received post the Cambridge Analytica scandal, the company has undertaken various steps to increase its credibility in terms of data use. Facial recognition technology is an important aspect in the broader unbiased AI space.
The focal point is to tame and train the algorithms accordingly. Improving ML capabilities on a dynamic basis will aid the process. Government regulations, transparency and accountability of the entities leveraging facial recognition technology is anticipated to make the transition from biased to unbiased AI smoother.
Zacks Rank & Key Picks
Currently, Microsoft carries a Zacks Rank #3 (Hold).
Better-ranked stocks in the broader technology sector are Western Digital (WDC - Free Report) and Mellanox (MLNX - Free Report) , both sporting a Zacks Rank #1 (Strong Buy). You can see the complete list of today’s Zacks #1 Rank stocks here.
The projected long-term earnings growth rate for Western Digital and Mellanox are 19%, and 15%, respectively.
Wall Street’s Next Amazon
Zacks EVP Kevin Matras believes this familiar stock has only just begun its climb to become one of the greatest investments of all time. It’s a once-in-a-generation opportunity to invest in pure genius.
Click for details >>