How do you Solve a Problem Like Big Data?

This post is not about how to analyze big data – it is about impact and implications for laws, business and our society.  How do we ensure that our increasing reliance on data, algorithms and AI does not come at a cost?

Below, I include some great articles on the topic.  Please read on, and would love to hear your thoughts via the comments section below.

Bloomberg: Battling the Tyranny of Big Data

Mark Buchanan explores recent research and efforts, such as the Open Algorithms Projects; he says we must control how data are used, second, open up by making it more widely available, then re-balance power between companies and individuals.

Mashable: We put too much Trust in Algorithms, and it’s Hurting our Most Vulnerable

Algorithms have gone wild in Ariel Bogle’s piece; they incorrectly label welfare recipients as cheats, and blacks, recidivist risks, give rise to “mathwashing.”

Huffington Post: We need to know the Algorithms the Government uses to make important Decisions about Us

Writer Nick Diakopoulos, Fellow at Columbia Tow Center; Assistant Professor of Journalism, University of Maryland, cites similar issues, and shares a case study in transparency in which he “guided students in submitting FOIA requests to each of the 50 states. We asked for documents, mathematical descriptions, data, validation assessments, contracts and source code related to algorithms used in criminal justice, such as for parole and probation, bail or sentencing decisions.”

PHYS ORG: Opinion – Should Algorithms be Regulated?

Offers a point vs. counterpoint; Markus Ehrenmann of Swisscom says “Yes.” Mouloud Dey of SAS says “No.”

Seth Godin’s blog: The Candy Diet

The marketing guru says that algorithms are dumbing down media.

The Drum: In the Post Truth Era, the Quest to Surface Credible Content has only Just Begun

Lisa Lacy reports that Google amended its algorithm to combat holocaust deniers; but if manual intervention is needed for high profile fails, but what about other important issues that don’t get as much attention?

Foundation for Economic Education: What Happens when your Boss is an Algorithm?

Cathy Reisenwitz offers a nice primer, and argues for making algorithms open source.

The New York Times: Data Could be the Next Tech Hot Button for Regulators

Steve Lohr voices concerns about the growing market power of big tech, and explains potential antitrust issues arising from their collection of data. He writes: “The European Commission and the British House of Lords both issued reports last year on digital “platform” companies that highlighted the essential role that data collection, analysis and distribution play in creating and shaping markets. And the Organization for Economic Cooperation and Development held a meeting in November to explore the subject, “Big Data: Bringing Competition Policy to the Digital Era.”

Fake News, AI and Bias, Oh My!

First published on DigiDig

There have been quite a few relevant articles recently, so I thought I would write this post which includes links and brief summaries.

Real Interest in Fake News

Fake news continues to be a hot story that shows what can go wrong when algorithms choose what we read. I covered the topic here blog when it first broke a few weeks ago.

Ryan Holmes of HootSuite wrote for the Observer in The Problem Isn’t Fake News – it’s Bad Algorithms: “As algorithms mature, growing more complex and pulling from a deeper graph of our past behavior, we increasingly see only what we want to see… More dangerous than fake news, however, is all the real news that we don’t see. For many people, Facebook, Twitter and other channels are the primary… place they get their news. By design, network algorithms ensure you receive more and more stories and posts that confirm your existing impression of the world and fewer that challenge it. Overtime, we end up in a “filter bubble”‘

Writing for DigiDay, Lucia Moses explained Why Top Publishers are Still Stuck Distributing Fake News. It is not just about news feed algorithms, but also involves the “intelligence” behind automatically-served programmatic ads (native ads can look like real articles). She shares an example in which the NY Times displayed a fake news ad next to their real story on fake news (got that?).

Algorithms and Bias

One of the primary concerns about algorithms relates to bias. How do biases infect data-driven computations? And in which ways do programs discriminate?

Kristian Hammond answers the first question in the TechCrunch story 5 Unexpected Sources of Bias in AI. He ponders whether bias is a bug or feature, and says: “Not only are very few intelligent systems genuinely unbiased, but there are multiple sources for bias… the data we use to train systems, our interactions with them in the ‘wild,’ emergent bias, similarity bias and the bias of conflicting goals. Most of these sources go unnoticed. But as we build and deploy intelligent systems, it is vital to understand them so we can design with awareness and hopefully avoid potential problems.”

Alvin Chang writes in Vox about How the Internet Keeps Poor People in Poor Neighborhoods. He shares an example of a Facebook ad that violates the Fair Housing Act by excluding certain users from seeing it. This is blatant, but algorithmic discrimination can be a lot more subtle, and thus harder to root out, he explains.

Artificial Intelligence for Dummies

If you are new to algorithms and AI, you might want to read this Digital Trends story, which breaks down the differences between machine learning, AI, neural networks, etc.

 

AI and Algorithms in the News

board-1364652_1920Cross-posted on DigiDig

I have been carefully watching for stories about the growing influence of technology in our lives, and sharing links with the DigiDig team via Toni Muzi Falcone. We discussed turning these into a weekly digest or DigiDigest (I truly hope that puns and weak humor attempts are not lost in translation; otherwise my writing tenure will be short here).

So, without further adieu, I list three very relevant stories.

Mother Nature’s Network – How Algorithms Influence us Every Day

This article describes ways in which algorithms help and hurt us. Cory Rosenberg worries that technology reduces our interests, backgrounds and behaviors to a number, and quotes Michele Willson of Curtin University in Perth, Australia: “Time, bodies, friendships, transactions, sexual preferences, ethnicity, places and spaces are all translated into data for manipulation and storage within a technical system or systems. On that basis alone, questions can be posed as to… how people see and understand their environment and their relations.”
Cory asks: “But isn’t the thought of humans as data an affront to our uniqueness?”
The article share examples of bias, e.g. “Google’s online advertising system shows ads for high-income jobs to males much more so than to females,” and examples of algorithms that help, like in “understanding” preferences in music and dating.

Mashable – Online Shopping Algorithms have us in a Decision Rut

Lance Ulanoff sees a “fundamental flaw in the technology designed to serve up things we might like. They are based entirely on past choices and activities and leave zero room for improvisation and unpredictability.” He bemoans the loss of serendipity in shopping recommendations, for example. It’s a timely topic in the US, as Cyber Monday was yesterday and the holidays are looming.

Lance writes: “If we continue to follow the choices made for us on social, services, subscription and retail sites, we will all soon be living a very vanilla life. Our friends will be the same kinds of people, our social feeds will offer just one point of view and our gift-giving will surprise no one.It is time to stand up and say, ‘You don’t know me.'”

Seems Lance would disagree with Cory’s view on the technology’s benefits.

Quanta Magazine – How to Force our Machines to Play Fair

Quanta writer Kevin Hartnett interviews author and Microsoft Distinguished Scientists Cynthia Dwork, who pioneered ideas behind “differential privacy.” She is now taking on fairness in algorithm design.

Cynthia says: “algorithms… could affect individuals’ options in life.. to determine what kind of advertisements to show people. We may not be used to thinking of ads as great determiners of our options in life. But what people get exposed to has an impact on them.”

She explores individual vs. group fairness and introduces the idea of “fair affirmative action.” Dwork would love to find a metric or way to ensure that “similar people [get] treated similarly,” but concludes that it is a thorny problem that people must first come to terms with before training computers to make these judgments.

In Defense of “Fake News”

More people are wondering about the weird crap that mysteriously appears in their news

Is News Today too Much Like the Magic 8 Ball?

Is News Today too Much Like the Magic 8 Ball?

feeds. How much is fake news? Did disinformation tilt an election? What are Google and Facebook going to do to clean up the mess?

You could almost hear the entire PR industry shifting uncomfortably amidst the backlash. I mean, crafting news (that some might call fake, or at least a stretch) is our stock in trade. We package propaganda as newsworthy information and sell it to the media; and, increasingly publish directly to the Web and social networks.

I understand that the fuss is more about blatant lies, not the average press release. But it highlights the challenges of determining what is newsworthy and true; a role that is increasingly being taken on by algorithms.

The Web and social media gave us all ways to easily share and spread information. This can include rumor, conjecture, commercial information, news, and yes, slander and outright lies.

I would never defend the last two; but will fight for our right to issue press releases, and traffic in other kinds of info. Any good system needs to be able to deal with all of this, i.e. anticipate some BS and surface the most credible and significant information, whether via the wisdom of the crowds, programs or a combination.

It is naïve to think that a publication, editors, or algorithms (which of course are written by humans) can present news without bias. The journalistic piece you just wrote might be pristine, free of opinion; but the very act of deciding which stories to feature shows partiality.

That said, the social networking platforms where more of us are getting news can do a much better job of separating the wheat from the chaff. I thought I’d share some of the great stories I’ve seen about the controversy and takeaways from each.

TechCrunch – How Facebook can Escape the Echo Chamber

Anna Escher says “Facebook is hiding behind its [position that] ‘we’re a tech company, not a media company’ … For such an influential platform that preaches social responsibility and prioritizes user experience, it’s irresponsible …”

She recommends that they bring journalists into the process, remove the influence of engagement on news selection during elections, and expand Trending Topics to show a greater diversity of political stories – not just the ones that are the most popular.

Tim O’Reilly – Media in the Age of Algorithms

Tim’s exhaustive Medium piece looks at all sides. He rails against “operating from an out-of-date map of the world [in which] algorithms are overseen by humans who intervene in specific cases to compensate for their mistakes,“ and says:

“Google has long demonstrated that you can help guide people to better results without preventing anyone’s free speech… They do this without actually making judgments about the actual content of the page. The ‘truth signal’ is in the metadata, not the data.”

Tim makes an analogy between news algorithms and airplanes “Designing an effective algorithm for search or the newsfeed has more in common with designing an airplane so it flies… than with deciding where that airplane flies.”

He cited an example from the history of aircraft design. While it’s impossible to build a plane that doesn’t suffer from cracks and fatigue… “the right approach … kept them from propagating so far that they led to catastrophic failure. That is also Facebook’s challenge.”

Nieman Lab – It’s Time to Reimagine the Role of a Public Editor

Mike Ananny writes about the public editor’s role, and the challenges they face in the increasingly tech-driven environment. He writes:

“Today, it is harder to say where newsrooms stop and audiences begin. Public editors still need to look after the public interest, hold powerful forces accountable, and explain to audiences how and why journalism works as it does — but to do so they need to speak and shape a new language of news platform ethics.”

He asks “Will the public editor have access to Facebook’s software engineers and News Feed algorithms, as she does to Times journalists and editorial decisions?” and says:

“… public editors must speak a new language of platform ethics that is part professional journalism, part technology design, all public values. This means a public editor who can hold accountable a new mix of online journalists, social media companies, algorithm engineers, and fragmented audiences — who can explain to readers what this mix is and why it matters.”

Latest PR Gambit: Publishing on Platforms

Back in the day (“the day” being about 10 years ago), we had a simple message for PR shoe-737084_1920clients who wanted to get in on the social media and blogging action.

It was: “Go forth and blog too. Master the channels that are accessible to all.” Those who took the time to produce quality content, nurture social communities and post consistently saw their online influence grow.

Now, the open web is being challenged by the growth of social networking platforms. They’re places we go to connect, and get entertained and informed. Their news clout is growing, as the networks are increasingly publishers and aggregators of content. The social networks reach vast audiences with precise targeting – compelling attributes for marketers.

In short, if you are in the news business or want to promote your own, you are missing out if you are not on Facebook, LinkedIn, Twitter, etc.

But there are a number of challenges along the way. It takes PR out of our media-centric comfort zones. It’s not obvious how to use social networking channels to accomplish your goals, which generally include coverage KPIs.

Sure, many in PR have jumped on the social media and content marketing bandwagons. We can handle Tweeting and blogging quite well. But getting your news seen and covered or appreciated by the right audiences, especially if your profile does not already have umpteen million friends/followers, is another matter.

Success generally requires a combination of paid and organic promotion as well as an understanding of the algorithms, those wonky programs that determine what appears in our news feeds. But they are black boxes and constantly changing. Plus, ad options may be unfamiliar, and they’re also moving targets.

How does one figure this all out? Listen, read, and more important, experiment. Dip your toes in. Test, validate, then repeat.

Reading this blog is a good start, as it offers commentary, articles about best practices and links to the right resources. The networks can be opaque, when it comes to specifics about their algorithms – but they do inform about changes and make recommendations.

In short, there are no pat answers, although one could invoke advice similar to the words at the beginning of the article: go forth and publish on Facebook (for example). Learn about the secrets of shareable content and how to get into the news feed.

I’ll close with an example from the world of politics, which seems fitting since the election has been front and center. It’s an article that ran awhile back in the NY Times Sunday magazine.

What do you think? Could a similar approach work beyond the field of politics? What ideas does this give you for PR? See the link and excerpts below, and please share your comments.

Inside Facebook’s… Political Media Machine
[Facebook’s] algorithms have their pick of text, photos and video produced and posted by established media organizations… But there’s also a new and distinctive sort of operation that has become hard to miss: political news and advocacy pages made specifically for Facebook, uniquely positioned and cleverly engineered to reach audiences exclusively in the context of the news feed…

These are news sources that essentially do not exist outside of Facebook… cumulatively, their audience is gigantic: tens of millions of people. On Facebook, they rival the reach of their better-funded counterparts in the political media…

But they are, perhaps, the purest expression of Facebook’s design and of the incentives coded into its algorithm — a system that has already reshaped the web…
Truly Facebook-native political pages have begun to create and refine a new approach to political news…. The point is to get [users] to share the post that’s right in front of them. Everything else is secondary.

DigiDig Studies the Impact of AI and Algorithms on Society

banner-1571986_1920Full story is on Flack’s Revenge

I’ve had many great conversations with my friend and PR authority Toni Muzi Falcone about the impact of technology on the field and society at large. Recently he told me about a new effort that he helped conceive – DigiDig – a website and citizen-led group dedicated to studying this area.

They started in Italy (the website for now is almost entirely in Italian) and have an international focus. I asked Toni to tell me more, and he shared the following:

“DigiDig questions the algorithmic society. It is a start-up community, launched on October 9, 2016, of some 150 Italian digitally active and prominent citizens wishing to better understand, discuss and raise awareness of peers on issues related to ‘user power’ vis-à-vis XXI century global robber barons.

The intention is to connect with the many similar or analog groups that populate the global digital space. Our immediate focus includes Brussels and New York, but we also wish to dialog with New Zealand, Kazakhstan and Namibia.

Its promoters are academics, intellectuals, journalists, managers, lobbyists, elected officials, communicators, writers, sociologists, and polemicists.

Following two open house sessions in Rome and Milano last June, a coordinating committee of six was formed and a web space just opened a few days ago containing a shared ‘manifesto’ plus opinions and comments in Italian and English.

The manifesto, titled: ‘the algorithm as a technology of freedom?’, defines its main issue as

(…..) the active and critical observation of the true nature of the global process reorganizing social and economic life, focused on the development and exchange of cognitive products of artificial intelligence.

As algorithms simplify digital procedures as well as the automation of humanity’s most delicate and discretionary activities, we cannot accept that such process proceeds in disrespect of the elementary rules of transparency, information and access to participation to its decision-making processes and operational standards.

If it is true that –as often affirmed by creators, shareholders and executives of those global groups – we are in fact confronted with a new ‘public sphere and/or space’ (and we very much believe it is so) – we also insist that the mechanisms creating new alphabets, social structures and determining influences over individual choices, need to be understandable, shared, socially negotiable and integrated.(…).

No membership fees, but requests for contributions via PayPal at info@digidig.it

Breaking into Facebook’s News Feed: 3 Stories, 9 Tips

Facebook has made quite a few changes to its algorithm and news feed in recent months, as has been news-1592592_1280chronicled on this blog.  Digiday said that some publishers are responding by focusing more efforts on SEO.

But where does this leave brands and marketers who want to target Facebook users with news and content?

You need to change with times. These days, your content should be informative, relevant and entertaining – it helps if the topics resonate with your friends and family.

There were quite a few good posts that recommended strategies in light of the updates.  Here were three that stood out, and three tips for each.

In A Publisher’s Guide to Facebook’s News Feed Updates, the Newswhip blog shared these tips:

  1. Focus on organic reach and stories shared by actual users vs. brand pages
  2. Use engagement metrics to inform strategy and content creation
  3. Stay attuned to what interests your readers and work hard to serve a niche audience

The same blog follow up with more good advice: How to Adapt to Facebook’s “Personally Informative” News Feed. It offered a helpful pointer to how Facebook defines Personally Informative. This means staying in tune with audience interests. Newswhip recommends:

  1. Creating an RSS aggregator featuring the news sources favored by your desired audience
  2. Building your personal brand – the changes favor peer-to-peer sharing
  3. Being genuine, avoiding clickbait and deception

To the last point, Facebook’s more recent changes target and penalize click bait.  The Hootsuite blog featured a story on How to Get Clicks without resorting to Clickbait.  It recommends:

  1. Be accurate, the headline shouldn’t promise more than the content delivers
  2. Create an emotional connection
  3. Take the time and care to craft an effective headline

 

Steal this News Feed (How to get Into Facebook Trending)

I don’t typically write about reverse engineering news feeds. This blog is about hacking the feed in a hacker-1500899_1280figurative sense; i.e. boosting the odds that your news gets featured in the social networks that dominate our attention these days. It’s less about black hat, more smart marketing and communications.

But I thought I’d share a story about the actual hacking of algorithms. In Quartz, David Gershgorn wrote that Stealing an AI Algorithm and its data is a “high school-level exercise.”  He wrote:

Researchers have shown that given access to only an API, a way to remotely use software without having it on your computer, it’s possible to reverse-engineer machine learning algorithms with up to 99% accuracy. Google, Amazon, and Microsoft allow developers to either upload their algorithms to their cloud or use the cloud company’s proprietary AI algorithms, both of which are accessed through APIs.

The article explained how you can crack the algorithm’s logic by sending queries, and evaluating the answers:

Think about making a call to a machine learning API as texting a friend for fashion advice. Now imagine you were to send your friend thousands of messages…  After driving your friend insane, you would get a pretty clear idea of their fashion sense, and how they would pick clothes given your wardrobe. That’s the basis of the attack.

However, anyone who wants to hack Facebook’s news feed would not benefit from this approach, which relies on the availability of an API that’s accessible to developers in the cloud.

So, what about figurative hacking? As this recent NiemanLab piece relates (it also references Gershgorn) Almost No one Knows How Facebook Trending Algorithm Works (But Here’s an Idea). Joseph Lichterman wrote:

Trending now… features broad topics surfaced by the algorithm. According to Facebook’s guidelines, the engineers overseeing Trending are “responsible for accepting all algorithmically detected topics that reflect real-world events.”

Based on some sniffing around, he determined that these thing can help you get into Facebook Trending:

  • Make sure your content includes keywords or hash tags that are trending
  • Don’t spam (Facebook detracts for frequent posting)

What do you think? I’ll be sharing many more tips about how to optimize your news for the Facebook news feed in an upcoming post.

Don’t Shoot Me, I’m Only the Headline Writer!

You may recall Elton John’s album “Don’t Shoot Me, I’m Only the Piano Player.” That’s what I thought piano-1589154_1280of when I heard about Facebook’s move to lessen the role of people in Trending topics.

As reported in the Washington Post: “Facebook just greatly diminished the role that human beings will play in the platform’s Trending topics bar, announcing … that actual people will no longer write topic descriptions for the site.  [This] comes months after the company faced an unusually high level of scrutiny for alleged political bias in its Trending feature. Humans will serve a janitorial role in the process, while the algorithms take more control.”

It is an interesting state of affairs when algorithms are deemed to be more unbiased than people – and we are serving the machine in a “janitorial role” (of course, Facebook did not come out and say this as I have – another article attributed the change to the need for scalability).

It seems clear, however that they are still smarting from the bias accusations. Perhaps Facebook’s trying to counter this by giving machines, which we think of as logical, more of a role.

There was a great article in Time magazine about the danger of placing too much trust in algorithms. Rana Foroohar wrote about Cathy O’Neil’s new book Weapons of Math Destruction, which highlights the growing role of algorithms in everything from job performance evaluations, to grading teachers, credit decisions, etc.  They determine which ads we see, and increasingly point us to (and describe) news topics.

Rana writes:

“The Big Data algorithms that sort us into piles of “worthy” and “unworthy” are mostly opaque and unregulated, not to mention generated (and used) by large multinational firms with huge lobbying power to keep it that way.”

How I Stopped Worrying and Learned to Love the Algorithm

The latest Facebook algorithm changes are in, and the tally is Facebook 1, headline writers 0.  As was dont-panic-1067044_1920widely reported over the past few days (e.g., see this TechCrunch piece), the social network has taken measures to reduce the number of click bait stories in our news feeds.

They’re trying to improve the user experience, by studying which types of stories people bounce from and coming up with a formula that flags the content (and marks the source as a click baiter).

Apparently this has a lot to do with the headline – does it withhold crucial information to tempt  curiosities, or over hype the article contents? Facebook offered these tips to help publishers comply. The changes are designed to reward quality and punish excesses of content creators.

Here, Facebook places us in an awkward position. I mean, who would argue for more click bait? The problem is that the types of things they’re now watching for are exactly the time-proven tactics that work, i.e. draw the user in. If the headline doesn’t pull, the shitting thing doesn’t get read.

The devil is in the details – I am not a believer in hypey headlines, and promising more than articles deliver. But hype is in the eye of the beholder, or now the algorithm. I used their sniff test while reading the esteemed NY Times, some of their stories could take a hit according to the screening logic.

Tell me please, exactly how a computer is supposed to make these judgments at scale?

It all gets back to what I was saying earlier, about fighting excesses.  When more brands are plying more content, you get lots of listicles, crappy info graphics and irritating come-ons that are too tempting to resist.  Inevitably, quality declines. (See my post about Open Spaces Marketing to learn how to avoid this trap).

Algorithm writers have been trying to stay ahead of the content deluge for years, guiding users to the higher quality stuff. Recall Google’s Panda update in 2011 (see this Search Engine Land piece) to smack down content mills. Few mourn the decline of those sites.

Looking at it this way, these companies are our friends, and doing us all a service.

As I like to say, if you live by the algorithm, you can die by it too.

As Google says, and as I imagine Facebook would agree, write great content for people not algorithms, and you’ll do just fine (OK, well we may need to rethink our headlines too).