Is there a fundamental flaw in AI implementation, as @jrossCISR suggests in her latest article for Sloan Management Review? She and her colleagues have been studying how companies insert value-adding AI algorithms into their processes. A critical success factor for the effective use of AI algorithms (or what we used to call expert systems) is the ability to partner smart machines with smart people, and this calls for changes in working practices and human skills.
As an example of helping people to use probabilistic output to guide business actions, Ross uses the example of smart recruitment.
But what’s the next step when a recruiter learns from an AI application that a job candidate has a 50% likelihood of being a good fit for a particular opening?
Let's unpack this. The AI application indicates that at this point in the process, given the information we currently have about the candidate, we have a low confidence in predicting the performance of this candidate on the job. Unless we just toss a coin and hope for the best, obviously the next step is to try and obtain more information and insight about the candidate.
But which information is most relevant? An AI application (guided by expert recruiters) should be able to identify the most efficient path to reaching the desired level of confidence. What are the main reasons for our uncertainty about this candidate, and what extra information would make the most difference?
Simplistic decision support assumes you only have one shot at making a decision. The expert system makes a prognostication, and then the human accepts or overrules its advice.
But in the real world, decision-making is often a more extended process. So the recruiter should be able to ask the AI application some follow-up questions. What if we bring the candidate in for another interview? What if we run some aptitude tests? How much difference would each of these options make to our confidence level?
When recruiting people for a given job, it is not just that the recruiters don't know enough about the candidate, they also may not have much detail about the requirements of the job. Exactly what challenges will the successful candidate face, and how will they interact with the rest of the team? So instead of shortlisting the candidates that score most highly on a given set of measures, it may be more helpful to shortlist candidates with a range of different strengths and weaknesses, as this will allow interviewers to creatively imagine how each will perform. So there are a lot more probabilistic calculations we could get the algorithms to perform, if we can feed enough historical data into the machine learning hopper.
Ross sees the true value of machine learning applications to be augmenting intelligence - helping people accomplish something. This means an effective collaboration between one or more people and one or more algorithms. Or what I call organizational intelligence.
Jeanne Ross, The Fundamental Flaw in AI Implementation
(Sloan Management Review, 14 July 2017)
Retailers have long used fragrances to affect the customer in-store experience. See for example Air/Aroma
So perhaps we can use smell to alert consumers to dodgy websites? An artist and graphic designer, Leanne Wijnsma, has built what is basically an air-defreshener: a hexagonal resin block with a perfume reservoir inside, which connects over Wi-Fi to your computer. When it notices a possible data leak (like the user connecting to an unsecured Wi-Fi network, or browsing a webpage over an unsecure connection) — puff! It releases the smell of data
.James Vincent, What does a data leak smell like? This little device lets you find out (Verge, 31 Aug 2017)
That's all very well, but it only sniffs out the most obvious risks. If you want to smell the actual data leak, you'd need a device that released a data leak fragrance when (or perhaps I should say whenever) your employer or favourite online retailer is hacked. Or maybe a device that sniffed around a corporate website looking for vulnerabilities ...
I'm sure my regular readers don't need me to spell out the flaws in that idea.
Related postsPax Technica - On Risk and Security
(November 2017)UK Retail Data Breaches
On Friday, Transport for London (TfL) declared that Uber was not fit and proper to hold a private hire operator licence. Uber's current licence expires next week. However, Uber can continue to operate in London until any appeal processes have been exhausted. (TFL Press Release, 22 September 2017
By Saturday afternoon, a petition in Uber's favour had raised half a million signatures. Uber seems to put more energy into campaigning against evil regulators than into operating within the regulations, and was evidently already prepared for this fight. (You don't send out messages to millions of customers at the drop of a hat without a bit of forward planning.) As Emine Saner writes,
"Calling for better legislation certainly is not as exciting as a glossy app, or whipped-up social media reaction, but it may make your trip home safer – and would be a better use of online petitions."
The protests follow a number of well-worn arguments
- Many users of the Uber service (especially young women) have become dependent on a cheap, convenient and supposedly safer alternative to public transport and expensive taxis.
- Many drivers have borrowed heavily to invest in the Uber business model, and fear being thrown into penury.
- This is an anti-competitive and technologically backward move, prompted by entrenched interests. And as TfL is itself a transport operator, it is not appropriate that TfL should regulate its competitors.
None of these arguments can be taken completely at face value.
- It is true that many women believe the Uber model is safer than the alternatives; however, some women have been raped, and other women have had extremely scary experiences. Uber is accused of failing to carry out proper checks, and failing to report serious incidents.
- Uber service is cheap not only because it cuts costs and exploits its drivers, but also because it is subsidized by Uber investors. This looks suspiciously like predatory pricing rather than fair competition. Analysts such as Izabella Kaminska argue that Uber will only become profitable when it has driven its competitors out of business, at which point it will be able to increase its prices. Like much of Silicon Valley, it appears to operate according to the Peter Thiel anti-competition playbook. Even Steve Bannon has been heard arguing for closer regulation of what are effectively monopoly platforms.
- Technology companies such as Uber sometimes describe themselves as "disruptive". While it is true that disruptions sometimes yield socioeconomic benefits, the belief that disruption is always good for competition is based on ideology rather than evidence. Regulation is generally opposed to disruption.
- And as Stephen Bush points out, it's not as a digital start-up company that Uber has fallen foul of regulations, but as an old fashioned minicab operator. (As John Bull explains, Uber London is just a minicab company; the app is operated by Uber BV in the Netherlands. This corporate separation helps Uber to finesse both regulation and tax.) Persuading politicians and economists to see Uber as a shining example of technological progress is just "a very, very clever marketing trick".
I'm quoting Steve Bannon because I'm just amazed to find something I agree with him about. Regulating platforms is not the same as regulating regular companies, and the general art of regulation needs a kick up the proverbial. However, that is no reason to diss the current regulations or regulators, who are doing the best they can with insufficient regulatory mechanisms and resources. Experience from other cities shows that if Uber can't get its act together, there are plenty others that can.
John Bull, Understanding Uber: It’s Not About The App
(Reconnections 25 September 2017)
Stephen Bush, The right are defending Uber, because they don't really understand it
(New Statesman 22 September 2017)
Martin Farrer, Nadia Khomami et al, More than 500,000 sign petition to save Uber as firm fights London ban
(Guardian 23 September 2017)
Ryan Grim, Steve Bannon Wants Facebook and Google Regulated Like Utilities
(The Intercept, 27 July 2017)
Izabella Kaminska. For references see earlier post Uber Mathematics 2
Sam Levine,'There is life after Uber': what happens when cities ban the service?
(Guardian 23 September 2017)
Jason Murugesu, Night bus or black cab - what will save stranded Londoners post-Uber?
(New Statesman 22 September 2017)
Andrew Orlowski, Why Uber isn't the poster child for capitalism you wanted
(The Register, 26 September 2017)
Emine Saner, Will the end of Uber in London make women more or less safe?
(Guardian, 25 September 2017)
Related posts (with further references): Platform
What are the differentiating forces in the fast food sector? Stuart Lauchlan hears some contrasting opinions from a couple of industry leaders.
In the short term, those fast food outlets that offer digital experience and delivery may get some degree of competitive advantage by reaching more customers, with greater convenience. Denny Marie Post, CEO at Red Robin Gourmet Burgers, sees the expansion of third-party delivery services as a strategic priority. So from agility
But Lenny Comma, CEO of Jack in the Box, argues that this advantage will be short-lived. Longer-term competitive advantage will depend on the quality of the brand. So from assurance
Stuart Lauchlan, Digital and delivery – which ‘D’ matters most to the fast food industry? Two contrasting views
(Diginomica, 16 August 2017)
Related post: Reach, Richness, Agility and Assurance
More Recent Articles