Summary of #AIChat on September 6th, 2018

#AIChat is a 60 minute monthly Twitter chat on artificial intelligence.

Where my guest(s) help you make sense of the artificial intelligence world.

To give you artificial intelligence ideas to use at work. Or at home.

September’s guests? Professors Yannakakis and Togelius. The experts in artificial intelligence and gaming.

When you click on the link, you’ll have a recap of the chat:


Tagged ,

#AIChat Summary (Wed, Jul 4/18)

This is a summary of #AIChat (it’s a Twitter chat on artificial intelligence).

On Wednesday, July 4th, our guest was Roxana Nasoi.

We talked about artificial intelligence.

And it’s impact on the medical field.

Here’s the link to this conversation:

Tagged ,

Can Artificial Intelligence (AI) Protect Your Content from Hackers?

The Internet.

It’s the key for billions of people to watch your show.

If you’re using a secure cloud technology, someone will try to steal your content.

The problem? Hackers are

  1. targeting your vendors or your servers. To get the movies and TV shows before they’re released and
  2. betting that your vendors have a weak network security. That allows them to steal and hold your content for ransom.

Your goals?


  • disrupt attacks before they occur
  • manage risk and compliance issues and
  • make better use your security resources.

From an anti-piracy point of view, you can use the best technologies to guard your content at every step of the creative process. But these innocent devices will be used to hack your server:

  • bonded cell/wifi camera backs
  • smart phones on set
  • USB stick and
  • personal information on remote productions.

Keep in mind that not all your customers have the same content security needs.

The system must be flexible to work with customer’s needs without having to create a customized solution for each one.

And you’ll need an integrated analysis, a powerful user insights into what people are watching.

One solution?

You could use a premium security to protect your content, like white box cryptography. Or watermarking.

Or use an audience detection per view transaction company. That bridges the gap between the theatrical release and streaming services. For example, companies such as XCinex offers a per viewer ticketing model. Everyone in the room must have a ticket as if they’re going to a movie or concert. No ticket? No show.

But have you considered how artificial intelligence (AI) can help you?

By using

  • machine learning (ML) to analyze data to protect your IT systems. Machine learning can help you to identify the weak network elements or hosts where malware can exploit the weakness in the network. While scanning the network to identify and eliminate false positives and
  • behavior analytics. Instead of using a learned behavior approach (pre-learn or dynamic learning) over a period of time, AI can analyze network, application and user behavior to identify problems to prevent attacks.

Detecting threats to your servers is the main focus of AI and machine learning’s technology push.

Machine learning, advanced analytics and cutting edge technology will help companies track the data from their networks and Internet.

Big data will help you stay one step ahead of the hackers.

Before they can steal your content.

End Notes

Kadakia, Smit. “So, How Does Machine Learning Apply to Cyber Security?” ITSP Magazine (

Ng, David, Ryan Faughnder and Paresh Dave. “Cyberattacks once again roil Hollywood but can anything be done about it?” May 23rd, 2017

Oggel, Peter. “Mastering security to transform your media business.” (, July 13, 2017

\Suciu, Peter. “OTT. The Challenges of Infrastructure: ‘Quality of Service’ not guaranteed.”TV Technology (, July 21, 2017

Thomas, Brian E. “Why AI is crucial to cyber security.” June 16, 2017

Tribbey, Chris. “Deep Inside The Internet Backbone.” (, July 17th, 2017

Vogel, Beth. “How Does Your Digital Transformation Effort Stack Up” May 3, 2017

Wong, Steve. “Hacking Hollywood: How Safe Is Your Data?”




Tagged ,

Summary of #AIChat with @ConnektInc (Tuesday, April 24th, 2018)

On Twitter (@nickhtang), I have a monthly Twitter Chat devoted to talking about artificial intelligence.

For this month’s talk? My guest was Tripp Boyle, SVP @ConnektInc. We talked about artificial intelligence, television and voice assistants.

Q1) Why could TV be a catalyst for voice shopping?





Q2) What’s the role of AI?





Q3)Which types of TV programs have the biggest t-commerce potential?



Q4) Why are consumers still resisting from using a voice assistant to buy a product during or after a TV show?






In the News

I was interviewed by Cathy Hackl (Twitter: @CathyHackl) on my thoughts on artificial intelligence and how it could change reputation management and communications.

Here’s the link to her article:





Artificial Intelligence (AI) – Your Friend or Foe in the Entertainment Industry?

*** A thank you to Mia Dand, CEO of for editing this article. ***

Artificial intelligence (AI) is currently used in the entertainment industry. Combined with machine learning (ML) it can help networks and others get their content seen by the right people(1). It can also suggest to viewers what to watch next based on their interests. The key? Analyzing the metadata that’s used in the search optimization, ad targeting and recommendations.

How does machine learning help program AI?


  • Using natural language processing to analyze the dialogue in the TV show. To creates a series of themes
  • Analyzing the dialogue between actors helps it and AI examine the mood and personality traits of the main characters which allows consumers to find specific scenes and episodes
  • Gaining a deeper understanding of the video content. An example? The personality traits being displayed by an actor.

For example, Amazon uses AI to analyze scenes, objects and faces. It helps the company figure out what the viewers are clicking on.

When media and entertainment (M&E) businesses use ML, it helps them:

  • build logo and brand recognition and
  • determine how audiences are reacting to the content. The data can be used to market, personalize and stop consumers from switching to another service.

As AI starts using voice search in devices, the pay TV industry uses it to improve the viewer experience. An example? Food Network launching a search capability on Amazon voice enabled devices (Amazon Echo, Echo Dot, Amazon Tap, Amazon Fire TV). The viewer asks the devices about Food Network programs, schedules, recipes and more.

In the advertising world, AI can be used to personalize the content for the customer and advertiser. By using content intelligence, AI can examine what the viewers are watching and when and provides the advertiser a one to one marketing approach to the viewer.

Advertisers should consider AI as their partner that helps them analyze the data to find a fast solution (2). The pros and cons of using AI?


  • AI can be used for targeting and helping discover people to target the ads. It includes bidding efficiently for the ads and
  • AI can be used to build B2B sales leads.


  • AI can’t answer questions that have a broad answer. For example, it can’t recommend which social media platform to use for an advertising campaign (there’s no consistent measurement across these platforms)
  • AI can’t analyze data from consumer surveys or media consumption data. Media planners need to define the data before AI analyzes it and
  • AI can’t write long ads.

In the film industry, AI helps a movie studio figure out whether or not to make the film by analyzing the screenplay. Or determines the box office potential of a film before it’s released by using a neutral network algorithm to analyze 30 years worth data (box office revenues, film budgets, audience demographics and casting information).

Here are two examples where AI was used to write a short screenplay or scene:

  • an AI named Benjamin wrote the short film Sunspring by examining hundreds of movie scripts from the 80’s and 90’s. This is the first time an entire screenplay has been written by AI.
  • In the film “It’s No Game”, AI was used to write David Hasselhoff’s line. The machine learning algorithm learned to create long sentences based on learning rules from a corpus (composed of dialogue taken from several collections of films and television series) of writing. Another algorithm was used to create short sentences from the words.

AI can automatically create an edited clip that can be sent to another company or given to a human editor to create the final clip. For instance, IBM Watson selected the 10 most usable moments from the horror movie “Morgan.” The human editor created the movie trailer from the clips. This was done in 24 hours instead of taking 10-30 days using a human editor.

As machine learning continues to evolve, media companies can use the technology to create other types of content such as series and movie reviews (AI can find the data from different reviews to see what the critics or consumers are saying about it on social media or online).

AI should be seen as a partner.

Continue reading

Tagged , ,

Content – It’s An Algorithm 

Hollywood and the TV industry spend a lot of money figuring out what people wanted to watch.

It was a hit and miss. Strong shows or movies could offset the weaker ones. Executives could give a potential TV show (eg: Seinfeld or Cheers) or another so-so movie a sequel. In hopes that it’ll find an audience.

Fortunately, once the technology companies like Amazon Video, Hulu, Netflix and YouTube arrived, content became an algorithm. An algorithm based on analytics, designed to help these companies gain insights into their customers.

The perfect example?


Using the:

A) 30 millions “plays” a day from its subscribers (when they pause, rewind and fast forward)

B) 3 million searches its subscribers does 

C) data its subscribers generate (when they watch the shows, on what devices) and

D) the viewers metadata that describes the talent, action, tone and genre

Netflix was able to use all the data to create original content. Original content such as “House of Cards.” “Orange is the New Black.” “The Crown.”

That’s the power of big data.

Big data that allows you to create hits.

Not misses.

NB. Need more proof that big data and the algorithm can create content and drive subscriber growth? Netflix’s recent quarter. Where the company gained 5 million international users. 2 million US users.


Bulygo, Zach “How Netflix Uses Analytics To Select Movies, Create Content and Make Multimillion Dollar Decisions”

Carr, David “Giving Viewers What They Want,” February 24th, 2013

Finimize “Netflix, too busy to Netflix n’ chill,” January 19th, 2017 email newsletter

TV….it’s an app


It used to be that the only option was buying a big tv screen, placing it in one’s living room or den, then everyone would gather around at a specific time to watch their favorite program.

Not anymore.

As Maria Rua Aguete, Research Director for Television Media at IHS wrote in her blog:

“US cable TV average revenue per user (ARPU) is $80…many will continue cutting the cord in favor of lower priced alternatives if they feel they’re overpaying for what they are getting.”

As a result, and with the advent of Roku, Chromecast, AppleTV, tablets and smartphones, television has become an app.

An app where everyone can watch when they want. And where they want.

According to #Digital90210, media consumption in apps grew 108% from Q2 of 2014 to Q2 2015. Applications that makes it easier for consumers to access the content and services that are important to them.

The perfect example? CBS News launching a new AppleTV app with no authentication required. With a focus on live video where users can browse while watching to find related video and video playlists. With the ability to bookmark on demand videos.

An interesting point. CBS News reports that the average viewing session is the longest when users use AppleTV versus other devices. 96 minutes per session.

In addition, according to Brian Wiser from Pivotal Research, viewing of TV programming on internet connected devices rose 62% in June 2016 vs a year ago when he analyzed Nielsen’s data. He also noted that there was a boost in viewing on gadgets (Roku, AppleTV and Chromecast) that pushed the total use of TV up 2.4% in the 18-49 age group. And accounted for 7.7% of all viewing.

From an addressable TV advertising perspective, Tim Hanlon, managing director of FTI’s Consulting Telecom, Media and Technology Practice, has some interesting insights:

  1. cord cutting and unbundling are affecting/pressuring the economics of MVPDS (MVPDS can’t raise their subscription rates to compensate for the decline in subscriber numbers) and
  2. the advantage of television’s scale, when it comes to advertising, is a “red herring” as marketers are comfortable with precision targeting because of digital advertising thanks to audience data and the IP-ization” of media channels.

In addition, as the television audience continues to fragment (eg: HBO Now, CBS All Access), any advertiser who wants to advertise on these channels can’t because no measuring device has emerged to measure the audiences.

Add that about 50% of online viewing occurs in ad free or ad light formats, advertisers are asking this question – will consumers gradually get used to viewing no ads? If they do, then what happens? How can they quantify the program’s being watched online?

The challenges, from an advertising point of view, remain. Without any audience statistics, advertisers are unable to determine where to place their ads.

From a customer POV? It’s never been more easier to watch content online when they want it.


Why you need niche OTT

The viewer controls the dial. Not you.
The media and entertainment world is changing.

Viewers want to see their content when and where they want. Without the annoying advertising. With no breaks in between episodes. Without the noise, expensive movie tickets and snacks. And without paying for an expensive cable bundle with useless channels they refuse to watch.

Even though Netflix, Hulu and Amazon Video are generating the most press for their product (and mainline TV networks belatedly enter the OTT scene), there’s still room for you to prosper amidst all the competition.

Operators need solutions to increasing video consumption

According to Cisco, video traffic is increasing. By 2019, 80% of all IP traffic crossing service provider networks will be video; 42% will be viewed online.

Operators will be challenged with managing this traffic and extending traditional TV services to new devices. One possible solution? Turn to content delivery networks (CDNs). This means
1) distributing local content caching and multi screen streaming platforms out towards the edge of the network. This in turn reduces operator bandwidth requirements for delivering loads of IP Video content while giving their customers multi screens experiences they want or
2) operators can use the CDNs to offer “whole sale” CDN services to their content provider partners and other B2B customers who will pay to distribute their content over CDN.

Either way, this brings an enhanced live and on demand TV experiences to a range of screens (connected TV, tablets and smartphones through a single cloud).

Niche OTT = future profits
Even though the operators have the pipes, they will still need content to prosper. This leads us to niche OTT. It’s growing rapidly. In the US, it was worth $4 billion in 2014. By 2018, $8 billion.
Even though there are a lot of problems in
A) integrating infrastructure components from different vendors
B) acquiring large amounts of premium content and

C) retaining subscribers

niche OTT has a lot of potential as consumers are willing to spend on the content they love and easily can’t get. Fans willing to pay for the merchandise, interact with the actors, create content and talk about the brand.

Three examples:
Crunchyroll. In 2013, the OTT had over 200,000 subscribers paying $7/month for premium, ad free access. By 2015, the channel grew to 700,000 subscribers with around $60 million in revenues from subscribers and ad revenues from the site’s 10 million registered users. That’s the reason why it secured an extra $22 million in funding.

NBCU’s digital chief Evan Shapiro. Not only is he launching Seeso but 9 others as well. A strategy similar to the expansion of base cable networks over the last several decades.

Klowdtv focus on Spanish programming:

Technological barriers getting cheaper by the day
First, the technical barriers to OTT cost less. It’s cheaper to outsource technology systems to others. The old way? Building a team that can manage content, delivery, transcoding, search, storage, management etc. Time consuming and expensive. The new way? Outsource this to others like Brightcove, MLS Advanced Media, thePlatform (Comcast), Ooyala (system integrators).
Second, tons of new entrants are selling and distributing long tail content for more narrow topics. These new entrants are coming from different backgrounds versus web companies like Netflix or Google. These multichannel video programming distributors (MVPDs) might have a major impact on the market. In terms of content delivery, each OTT provider’s content delivery strategy (IP peering, private caching programs, use of SSL, multi-CDN sourcing) is an extension of their business strategy, the key to achieving desired cost and quality goals.

Other considerations 

According to Ralf Jacob, President of Verizon Digital Media Services, some OTT providers platforms were designed with VOD in mind. Live video is the perfect opportunity to target advertising at individual viewers, rather than broad demographic segments. Be prepared for a 100x spike as viewers tune in at the same time. This means your servers must be ready to handle big surges in viewership. 

In an ever changing market, the viewer wants to see compelling content at their convenience. Even though a lot of viewers tend to watch Netflix or Amazon Video, there’s enough room in the marketplace to create your own OTT channel.
One which allows you to capture and profit from a niche audience.
An audience which is willing to pay for your content. Spread the word about your channel. And becomes a devoted fan of your product.

P.S. An update. For those thinking that the US cable industry will be DOA due to the cord cutting, SNL/S&P Global Market Intelligence has some interesting statistics:

P.P.S. Update number 2: from an online video platform technology POV, continued growth as these equipment providers continue to provide the glue for everything to work together:

—, Cisco Readies Shaw’s Content Delivery Network to Share Thousands of Video Sources Across Multiple Devices, January 6th, 2016 courtesy of @CiscoSVP Video tweet (

Frankel, Daniel Fierce Cable “Seeso boss NBCU will go creating niche SVOD services just like it did cable channels” January 11, 2016 (

Haynes, Megan StreamDaily TV “How niche is too niche?” November 27th, 2015 (

Hultgren, Kaylee Cablefax “OTT Services to Grow to + $8 Billion by 2018; Niche Services on the Rise” July 15, 2015 (

Jacob, Ralf M&E Journal “How to Power a Live Video Event”, April 12th, 2017 (


Taylor Swift – How OTTs Can Copy Her Branding and Succeed

OTT. It continues to grow.Everyday you see companies entering the market to try and capitalize on their content.

Fortunately, a lot of them are using the exact tactics Taylor Swift uses to capture the market (here’s the link to the article – I will be using the same title headings from Thank you for the idea.

1) Be willing to reinvent yourself.
A perfect example? CBS. Even though it took the company a while to study the issue, when the company saw the rise of Netflix, Hulu etc, they decided to create their own OTT app after figuring out that there was going to be a demand from it. 

2) Lead the conversation.
Netflix. The company that started it all. And continues to lead by example. When content providers decide that they’ll gradually release their content to you or create their own OTT? Create your own content like Orange Is the New Black. When you realize that it’s possible to use several vendors to store and transmit your content via the Internet? Create your own cloud based content service. Anything that creates buzz and keeps you ahead of your competition.

3) Own your channel through branded content.
A lot of people only think of Amazon as a place to buy goods. Until Amazon decided to get into the content game by creating the Transparent series. Not only was this a serious challenge to Netflix but it allowed Amazon to take the lead from Hulu, which is floundering. 
Through it’s branded content, Amazon essentially told the world “[h]i….we’re here to play. And we can create the content to do this.” 

4) Humanize your brand
With Netflix, it’s through Ted Sarandos, who patiently explains to the analysts the reasons why they’re entering a particular market or creating exclusive content. Ted’s the human face behind Netflix. The person who makes it work.

5) “Think different”
A lot of OTT are thinking different. Unfortunately, all they’re doing is using the same cable marketing and hoping that it’ll carry over into the OTT world. Even though some people might buy into this hype, a lot of them will go “meh.” Perfect examples? In Canada, CraveTV and shomi. No buzz. No outstanding content. 

Netflix thinks differently. By creating different forms of content (and constantly measuring who’s watching or clicking it), it gives the audience a reason to tune in. And to customize their viewing habits. For instance, if one’s watching a drama, Netflix will suggest other pairings (similar to wine and food) to see what piques their interest.

As OTT continues to evolve, it’s up to the companies to tell their clients their “story” on why they’re unique in the marketplace. When they fail to do so, they only have themselves to blame.

Tagged , , , ,