Enterprise hits and misses - scraping Cambridge Analytica for insights, and negotiating better AI projects

Profile picture for user jreed By Jon Reed March 27, 2018
This week: Extracting lessons from Cambridge Analytica's ethical morass. Plus: the art of negotiating AI projects, and minimizing "risk bias." Google and Uber face setbacks worth watching. And yes, loads of goofy and baffling whiffs.

Cheerful Chubby Man

Lead story - Cambridge Analytica - enterprise lessons and overseas fallout by Den Howlett and Derek du Preez

The biggest consumer tech story of the week was also the biggest enterprise story of the week. Den assessed the bitterness of grandiose ambitions shriveling in an ethical morass fallout in How the Cambridge Analytica numbers don't stack up.

Taking his own due diligence medicine, Den combed through the numbers and filings, before concluding:

Today, we can only speculate but it is reasonable to assume that Cambridge Analytica LLC was the recipient of far more by way of campaign funds than appears on the surface, most likely via Giles-Parscale. The only question remaining is confirmation of quantum and where those funds came from.

This all blows back on Facebook, and as Derek reports, the EU is not amused (Zuckerberg faces pressure from EU over data sharing furore and resists calls from UK government to give evidence). After citing the mounting disappointment over Zuckberberg's leadership and Facebook's market loss ($100 billion and counting), Derek finds a bigger lesson:

The vitriol around the debate needs to lead to sensible debate about what controls and safeguards can be put in place to limit scandals of a similar nature in future. It’s not that hard, really. Users don’t mind handing over their data if they know exactly what that data is being used for. And they want control in insight into how that data is being used.

Too many companies in the "freeconomy" have played it fast and loose here. That unscrupulous data party seems to be winding down. It leaves enterprises with an opening to pursue a white hat data relationship with customers - or they can be next in line for the regulatory stockade. Yes, citizens must be more vigilant. But as Den reminds in Cambridge Analytica - an object lesson in failed due diligence and deception, enterprise buyers need to raise their game also:

Once again, we have an example of a business where due diligence is useful because if your business is seeking to influence buying outcomes, then it’s useful to know who is pulling the strings before someone else finds out.

In short, if there are any doubts then dig, dig and dig again.

Happy children eating apple
Diginomica picks - my top three stories on diginomica this week

Vendor analysis, diginomica style. Here's my four top choices from our vendor coverage:

Jon's grab bag - Picking college basketball games is no longer the purview of gamblers, suckers and fools office pools. As Kurt explains in Google's March Madness can help demystify machine learning, AI buffs and algo hackers are getting in on the action. This has relevance beyond  the hard wood: "The same sorts of data selection, normalization and model testing that goes into picking winners of a game, can be applied to many other questions and data sets."

Jerry's got the feel good (and feel good about blockchain) story of the week in How blockchain is providing ‘proof of existence’ for the world’s 1.1 billion refugees, though we're not nearly up to 1.1 billion on this blockchain just yet. I wrote about thriving in the robotic age, Jessica wrote on Wi-Fi access and mobile experience at CERN, and: if you haven't checked our first e-book offering, appropriately dubbed a d•book, I highly recommend this Phil Wainewright all-about-XaaS buzzword de-mystifier and handy opus - free for sign up.

Best of the rest

Waiter suggesting a bottle of wine to a customer
Lead story - Enterprise Technology: 6 Steps for Negotiating Winning AI Deals via UpperEdge

myPOV: Most of UpperEdge's posts focus on the bread and butter of contract negotiations with the enterprise behemoths. This post had a topical twist. Some companies aren't ready to dive into AI deals, but it's never too early to hone your BS filter. One key point: put the AI objectives in the contract. And: define what failure means:

This includes not only a stated performance objective for the AI system but also a definition of what would constitute failure and the legal consequences thereof. For example, in a contract for the use of AI in production management, is the objective to improve performance or reduce specific problems? And what happens if the desired results are not achieved?

I know, it's mind-blowingly inconceivable that an AI project wouldn't achieve its desired result, but that's the necessary fallback of a savvy contract. Clarifying the vendor roadmap is another vital point; most companies will align with a vendor on AI projects, and that may prove wise or costly. Plenty to consider here; I liked that UpperEdge raised the issue of bias impacting output:

Who is accountable or liable for incorrect outputs in situations where your business depends on the accuracy of AI and how will those situations be handled?

Something tells me companies are charging ahead without a clear answer to that question.

Honorable mention 


Overworked businessman
Let's get this party started:

Someone at Microsoft has lost the plot, albeit in a charmingly unselfconscious way:

I got a kick out of the schlock-kings-of-autoplay-videos, ZDNet, running this one without irony:

Oh, and regarding Facebook, I'm not sure this is going to the most effective boycott:

This breathless foolishness from Buzzfeed got my dander up the most:

First off, Musk and Zuckerberg are duelin' bros of the digital enlightenment. There is a huge and painless PR payoff in Musk deleting his Facebook Pages, with no business consequences. I do think Facebook may face some real consequences here, but if Buzzfeed is right, and we've arrived at the point where we can't take meaningful action until our geewhiz spacedaddy Elon Musk hits the delete button, we're a lot worse off than I thought.

Looks like science might not save us either:

Brings to mind a very old line from Tom Lehrer's satirical classic:

"Once the rockets are up, who cares where they come down? That's not my department, " says Wernher Von Braun

Make me want to ride a horse to McDonald's. Or, hitch a rocket ride with a nutty flat-earther, or: go marry a tree.

Whoops, that tree is taken. See you next time...

If you find an #ensw piece that qualifies for hits and misses - in a good or bad way - let me know in the comments as Clive (almost) always does.

Most Enterprise hits and misses articles are selected from my curated @jonerpnewsfeed. 'myPOV' is borrowed with reluctant permission from the ubiquitous Ray Wang.

Image credit - Cheerful Chubby Man © RA Studio, Happy Children © Anna Omelchenko, Waiter Suggesting Bottle © Minerva Studiom, Overworked Businessman © Bloomua, Businessman Choosing Success or Failure Road © Creativa - all from Fotolia.com.

Disclosure - SAP, Oracle, Workday and Salesforce are diginomica premier partners as of this writing.