AI Talk: Steps, SmartNews and voice prints

August 16th, 2019 / By V. “Juggy” Jagannathan, PhD

This week’s AI Talk…

10,000 steps per day

That’s the mantra preached by the medical establishment as a gateway to good health. Last month, the Journal of the American Medical Association reviewed a study which lowered this lofty goal some. The study discussed in this article found significant benefit with fewer steps. What exactly was the benefit they were looking for? Answer: Did the exercise help in delaying death? Well, this study looked at the impact of walking on a group of 16,741 individuals with the average age of 72 and followed them over a four-year period. They found a correlation between number of steps walked and number of deaths recorded in this population over the study period. Median steps of 2,718, 4,363, 5,905 and 8,442 per day led to 275, 103, 77 and 49 deaths respectively. The message from the authors: If you are already walking 10,000 steps per day, continue doing that! The study, which only looked at mortality, concluded that walking more will not only impact mortality, but your quality of life. That I believe is good advice indeed.

Voice Print

This week, The Wall Street Journal published an article on voice prints and what they reveal about us. For starters, they are becoming a biometric which can be just as effective as a fingerprint! And, the applications seem boundless, from security to identifying scammers to clinical applications—a lot of companies have begun to mine this new treasure trove of data. Right in my backyard, CMU researcher Rita Singh is using voice to figure out a person’s skull structure, height, weight, etc. Discover Financial Services apparently managed to reduce its losses due to scamming using this tech. But what is particularly intriguing is its use in identifying potential conditions that a person may be suffering from, like depression, dementia and Alzheimer’s. Of course, there are privacy concerns, along with concern about misuse of the data. Some states have legislation requiring a person’s consent before using their voice as data, but it is unclear what one is consenting to.


SmartNews

I saw an article this week in Nanalyze about a Tokyo-based startup called SmartNews, which reached a lofty valuation of $1.1 billion this month! This company is a news aggregator—aggregation using AI from a multitude of source platforms to deliver news that is targeted to the user’s preference. Google News employs the same technology, but apparently this company is successfully competing with Google. It is also making good money providing this service. I am a news junky as you can imagine—particularly when it comes to tech and AI news—so maybe I should give this a whirl! AI is used in almost every aspect of such aggregation and it is interesting to note companies creeping up in this space.

Section 230

What is section 230? That was my question—and probably most other folks as well! It turns out President Trump has turned his attention to this oft misunderstood provision in the Communication Decency Act. MIT Technology Review has an excellent article explaining this legal corner and how it enabled the modern-day internet’s character and form. What exactly is section 230? Here is an excerpt from the article: “Section 230 of the Communications Decency Act of 1996 states that, with some exceptions, internet companies are not legally responsible for the content they host if it was published there by someone else…and, they are not obliged to host anything they don’t want to.” This provision has shielded communication platforms such as YouTube, Twitter, Facebook and others while allowing them to take down content that they deem objectionable. Therein lies the rub—the definition of objectionable content is left to the discretion of platforms. President Trump claims they have an anti-conservative bias, the Speaker of the House Nancy Pelosi claims the opposite. Misconceptions about this law abound and Congress is making noise about changing this section. What exactly should the changes be? No one has a clue!

I am always looking for feedback and if you would like me to cover a story, please let me know. “See something, say something”! Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is an AI Evangelist with four decades of experience in AI and Computer Science research.