Silicon Valley’s ‘ask for forgiveness, not permission’ attitude is wearing thin

Silicon Valley has often pushed the boundaries in pursuit of progress, but the it deserves everything it gets if it continues to try the patience of consumers and regulators with privacy.

‘It is easier to ask for forgiveness, than beg for permission’ is a common, if largely unattributable, phrase which seems to apply very well to the on-going position of Silicon Valley. It is certainly easier to act and face the consequences later, but it should not be right or allowed. This is the approach the internet giants are taking on a weekly basis, and someone will have to find the stomach and muscle to stop this abuse of power, influence and trust.

The most recent chapter in this on-going tale of deceit and betrayal concerns the voice assistants which are becoming increasingly popular with consumers around the world.

Apple is the latest company to test the will of the general public as it has now officially ended an internal process which is known as ‘grading’. In short, humans listen to Siri interactions with customers, transcribing the interaction in certain cases, to help improve the accuracy of the digital assistant.

“We know that customers have been concerned by recent reports of people listening to audio Siri recordings as part of our Siri quality evaluation process — which we call grading,” Apple said in a blog entry. “We heard their concerns, immediately suspended human grading of Siri requests and began a thorough review of our practices and policies. We’ve decided to make some changes to Siri as a result.”

Of course, it is perfectly reasonable for Apple to want to improve the performance of Siri, though it must ask for permission. This is the vital step in the process which Apple decided to leave out.

The new process will seek consent from users through an ‘opt-in’ system, making it compliant, while the default position for all Siri interactions will be to not store information. For those consumers who do opt-in to aid Apple in training Siri, the audio will only be transcribed and reviewed by permanent Apple employees.

This process should have been in-place prior to the ‘grading’ system being implemented. It is inconceivable that Apple did not realise this would break privacy regulations or breach the trust it has been offered by the customer. It decided not to tell the consumer or authorities this practice was in place. It muddied the waters to hide the practice. It lied to the user when it said it respects privacy principles and rights.

Apple acted irresponsibly, unethically and underhandedly. And there is almost no plausible explanation that it did so without knowledge and understanding of the potential impact of these actions. If it did not understand how or why this practice violated privacy principles or regulations, there must be an epidemic of incompetence spreading through the ranks at Cupertino.

What is worth noting is Apple is not alone; Google and Facebook are just as bad at misleading or lying to the user, breaking the trust which has been offered to these undeserving companies.

Google is currently under investigation for the same abuse of trust and privacy principles, this time for the Google Assistant.

“We have made it clear to Google’s representatives that essential requirements for the operation of the Google Assistant are currently not fulfilled,” said Johannes Caspar, Hamburg Commissioner for Data Protection and Freedom of Information. “This not only applies to the practice of transcribing, but to the overall processing of audio data generated by the operation of the language assistance system.”

The investigation from the Hamburg data protection authority has pressured Google into changing the way it trains its digital assistant. Earlier this month, Belgian news outlet VRT NWS revealed 0.2% of conversations with Google Assistant were being listened to by external contractors. At least one audio clip leaked to the news outlet included a couple’s address and personal information about their family.

Google has now said it has stopped the practice in the EU, but not necessarily elsewhere, and the Hamburg DPA has said it will have to seek permission from users before beginning anything remotely similar.

At the same regulator, Facebook has been dragged into the drama.

“In a special way, this also applies to Facebook Inc., where as part of the Facebook Messenger to improve the transcription function offered there a scheduled manual evaluation was not only the human-to-machine communication, but also the human-to-human communication,” said Caspar. “This is currently the subject of a separate investigation.”

Two weeks ago, reports emerged Facebook had hired external contractors to transcribe audio from calls made across the Messenger platform. Once again, users were not informed, while consent was not obtained, but what makes this incident even worse, is there does not appear to be any logical reason for Facebook to need this data.

The only reason we can see why Facebook would want this data to improve algorithms is to take the insight to feed the big-data, hyper-targeted advertising machine. However, this is a massive no-no and a significant (and illegal) breach of trust.

All of these examples are focused on transcription of audio data, though there are many other instances of privacy violations, and demonstrate the ‘easier to ask for forgiveness than permission attitude’ which has engulfed Silicon Valley.

We cannot believe there is any way these companies did not understand or comprehend these actions and practices were a breach of trust and potentially breaking privacy rules. These companies are run by incredibly smart and competent people. Recruitment drives are intense, offices and benefits are luxurious, and salaries are sky-high for a very good reason; Silicon Valley wants to attract the best and brightest talent around.

And it works. The likes of Google, Facebook and Apple have the most innovative engineers, data scientists who can spot the wood for the trees, the savviest businessmen, accountants who are hide-and-seek champions and the slipperiest lawyers. They consider and contemplate all potential gains and consequences from any initiative. We cannot believe there is any conceivable explanation as to why these incredibly intelligent people did not recognise these initiatives were either misleading, untransparent or non-compliant.

The days of appearing before a committee, cap in hand, begging for forgiveness with a promise it will never happen again cannot be allowed to continue. The judges, politicians and consumers who believe these privacy violations are done by accident are either incredibly naïve, absurdly short-sighted, woefully ill-informed or, quite frankly, moronic.

Silicon Valley must be forced to act responsible and ethically, because it clearly won’t do it on its own.

The smart speaker revolution isn’t quite here yet

A survey on voice assistant use in the US reveals the majority of consumers hardly ever use smart voice assistants and even when they do it’s on a smartphone.

The survey was conducted by digital commerce strategy firm Sumo Heavy. It chatted to around a thousand US punters to find out what their voice assistant habits are. The first thing it found was that 46% have never used  voice assistant and that 19% do so rarely, which probably means they did it once out of curiosity and concluded they didn’t fancy it.

sumo slide 1

So that means less than a third of US consumers use voice assistants with any regularity. Considering this sort of technology hit the mainstream when Siri was pre-integrated into the iPhone 4S back in 2011, it would be a stretch to say voice assistant adoption has exploded in the intervening 8 years.

The same could be said for smart speakers, which had their mainstream launch in 2014 with the launch of the first Amazon Echo. We have been led to believe they achieved near ubiquity in the subsequent 4-5 years but this survey begs to differ. By far the most common way of accessing voice assistants is smartphones, with smart speakers accounting for less than a fifth.

sumo slide 2

Unsurprisingly those people that do own smart speakers (Amazon is the clear market leader) use the voice assistant fairly frequently, with over half doing so at least once per week. Iphone users seem more into voice assistants than Android ones. Only 42% of regular voice assistant users ever buy stuff that way, but if they do it’s likely to domestic products or things like movie tickets.

Turns out real people sometimes hear what you say to smart speakers

The revelation that Amazon employs people to listen to voice recordings captured from its Echo devices has apparently surprised some people.

The scoop comes courtesy of Bloomberg and seems to have caught the public imagination, as it has been featured prominently by mainstream publications such as the Guardian and BBC News. Apparently Amazon employs thousands of people globally to help improve the voice recognition and general helpfulness of its smart speakers. That means they have to listen to real exchanges sometimes.

That’s it. Nothing more to see here folks. One extra bit of spice was added by the detail that sometimes workers use internal chatrooms to share funny audio files such as people singing in the shower. On a more serious note some of them reckon they’ve heard crimes being committed but were told it’s not their job to interfere.

Amazon sent Bloomberg a fairly generic response amounting to a justification of the necessity of human involvement in the AI and voice recognition process but stressing that nothing’s more important to it than privacy.

Bloomberg’s main issue seems to be that Amazon doesn’t make it explicit enough that another person may be able to listen into your private stuff through an Echo device. Surely anyone who knowingly installs and turns on a devices that is explicitly designed to listen to your voice at all times must be at least dimly aware that there may be someone else on the other end of the line, but even if they’re not it’s not obvious how explicit Amazon needs to be.

An underlying fact of life in the artificial intelligence era is that the development of AI relies on the input of as much ‘real life’ stuff as possible/ Only be experiencing loads of real interactions and scenarios can a machine learn to mimic them and participate in them. In case there is any remaining doubt, if you introduce a device into your house that is designed to listen at all times, that’s exactly what it will do.

Alexa turns to HERE to crack the car market

Amazon is collaborating with navigation platform HERE in order to get its Alexa voice UI into the car of the future.

The announcement was made at CES 2019 and involves the integration of Alexa into the HERE navigation and location platform, thus giving it a voice UI dimension. This seems pretty sensible as in-car infotainment systems are already too complex to be safely operated via a touch screen, meaning cars are the perfect setting for enhanced voice interactions.

“The in-vehicle user experience is rapidly changing, and automakers today have the opportunity to deliver the next generation of services that maximize the vehicle’s utility as the ultimate connected device and providing consumers with the user experience they expect,” said Edzard Overbeek, CEO of HERE Technologies, before pausing for breath. “Our work with Amazon will drive a truly differentiated and delightful user experience, from the home to the car, to where you want to go, and what you need to know.”

In a parallel announcement HERE launched a new version of its platform called HERE Navigation On Demand, which is positioned as ‘The world’s first SaaS navigation and connected service solution for vehicles’. It seems to use the core SaaS concept of allowing OEMs and their customers to easily cherry-pick the aspects of the suite their individual needs dictate.

“HERE Navigation On Demand is the reinvention of in-car navigation for the era of the connected vehicle,” said Overbeek. “Our solution gives automakers the agility and flexibility they need to deliver the most competitive navigation experiences on the market. Moreover, it provides them the freedom to create their own business models that support their unique strategies.”

“We’re thrilled to be working with HERE to integrate Alexa with its in-vehicle navigation software,” added Ned Curic, VP of Alexa Auto at Amazon. “Because Alexa is integrated directly into the experience, automakers using HERE Navigation On-Demand can easily provide customers with an intuitive, voice-first experience in the car, and provide richer, more useful voice interactions at home and on the go.”

The in-car infotainment platform was has been fermenting for some time but it could be set to escalate. Google announced a big partnership back in September of last year and presumably isn’t keen to share the dashboard with HERE and Alexa. Forcing OEMs to make a long-term commitment to one platform probably isn’t a good idea however, which is why HERE may have been clever to adopt the SaaS model.

Wifi mesh technology could kick-start the smart home

An emerging technology designed to create a seamless domestic connectivity experience could create a bunch of other opportunities, according to Qualcomm Atheros.

We spoke to Irvind Ghai, VP of Product Management at Qualcomm Atheros, at a briefing in London. After covering the recent announcements of some new 5G NR small cells and a collaboration with Facebook over its Terragraph FWA project, we moved onto wifi, which is one of the areas Ghai’s bit of Qualcomm focuses on.

One of the most interesting concepts we covered was wifi mesh, which involves installing multiple (typically three) wifi nodes in the house to extend the range of a router. Unlike current fixes such as wireline wifi extenders, a mesh has additional cleverness that enables your connected devices to dynamically hand over between nodes depending on which provides the best signal.

The really clever bit, however, lies on some of the ancillary stuff this technology enables. Of greatest interest to CSPs could be a radar-like ability to map the interior of the home, which enables localised responses to voice commands. For example you could say “lights on” when you’re in the kitchen and the smart home system would only turn the lights on in the kitchen.

In fact these sorts of systems can apparently support their own voice UI systems and, such is the precision of this domestic radar that it can also support things like gesture UI. On top of that it can detect when doors and windows are open, so it seems to offer lots of tools for CSPs to fashion into a compelling smart home proposition if they can just get their acts together.

Ghai told us that mesh products already account for 40% of the US domestic wifi market and pointed to vendors such as Plume (mesh nodes pictured above), which make small, unobtrusive nodes that can be discretely placed around the house. You can see some of Ghai’s slides below and if wifi mesh delivers as advertised it could be a significant technology for the development of the smart home.

Qualcomm wifi 1

Qualcomm wifi 2

Qualcomm wifi 3

Qualcomm wifi 4

Ericsson urges CSPs to be less touchy

Telecoms vendor Ericsson has been looking into how CSPs interact with their customers and reckons they take too long to get stuff done.

It’s a good day for CSP top tips and it seems likely that a healthy proportion of subscribers would agree that their interactions could be significantly improved. While operators have already identified customer service as a priority, it seems progress may not have been as rapid as they’d hoped.

According to Ericsson’s Consumer & IndustryLab Insight Report, a typical interaction between subscriber and CSP takes 2.2 attempts and 4.1 days to complete. While we may have become conditioned to such glacial progress when trying to get things done, there seems to be a clear opportunity for CSPs to pleasantly surprise their customers by not being rubbish.

“Consumers believe telecom service providers treat touchpoints like isolated interactions,” said Pernilla Jonsson, Head of Ericsson Consumer & IndustryLab. “Siloed focus means they miss the bigger picture. Interestingly, telecom service providers could leapfrog one-click and move from multiple-click to zero-touch by deploying future technologies in their customer offerings. The zero-touch customer experience report shows that zero-touch experiences are now an expectation of their customers.”

While it’s hardly surprising that a telecoms vendor should recommend their customers spend more money on new technology, Ericsson may also have a point. The proposed technological solution involves lashings of artificial intelligence and analytics. This mainly involves anticipating their every need and satisfying it in advance and the ‘zero touch’ thing seems to refer to alternative UIs such as voice.

AI audio is getting scary

Google is trying to make machines sound more human and that’s freaking people out.

Earlier this week Google demonstrated a cool new technology it’s working on called Duplex that is essentially an AI-powered automated voice system designed to enable more ‘natural’ conversations between machines and people. You can see the live demo below and click here for a bunch of audio clips showing how far along it is.

While there is clearly still a fair bit of fine tuning to be done, the inclusion of conversational furniture such as umms and ahs has unsettled some commentators, mainly on the grounds that it’s becoming hard to know if you’re speaking to a real person or not. While the whole point seems to be to make interacting with machines more smooth and intuitive, it seems we’ve hit a cognitive wall.

‘Google Grapples With ‘Horrifying’ Reaction to Uncanny AI Tech’, blurted Bloomberg. ‘Could Google’s creepy new AI push us to a tipping point?’ wailed the Washington Post. ‘Google Assistant’s new ability to call people creates some serious ethical issues’, moaned Mashable. ‘Google Should Have Thought About Duplex’s Ethical Issues Before Showing It Off’, fulminated Fortune. And then there’s this Twitter thread from a New York Times writer.

Hyperbolic headline-writing aside these are all good points. Google’s grand unveiling coincides with the broadcast of the second season of Westworld, a drama in which androids indistinguishable from humans rebel and decide to start calling the shots. And, of course, talk of AI (at least from this correspondent) is only ever one step away from references to The Terminator and The Matrix.

The above reactions to the demonstration of Duplex have forced Google to state that such interactions will always make it clear when you’re talking to a machine but it’s not yet clear exactly how. More significant, however, has been this timely reminder that not everyone embraces technological advancement as unconditionally as Silicon Valley and that AI seems to have already reached a level of sophistication that is ringing alarm bells.

And it’s not like Duplex is an isolated example. The NYT reports on findings that it’s possible to embed suggestions into recordings of music or spoken word such that smart speakers receive them as commands. The extra scary bit is that it’s possible to make these commands undetectable to regular punters.

Meanwhile Spotify has announced a new ‘Hate Content and Hateful Conduct Public Policy’, that is enforced by an automated monitoring tool called Spotify AudioWatch. This bit of AI is able to sift the lyrics of songs on the platform for stuff that goes against Spotify’s new policy, which you can read here.

On one hand we can all agree that horridness is bad and something needs to be done about it, on the other this is yet another example of algorithmic censorship. According to Billboard this facility is also being used to erase from history any artists that may have sung or rapped something horrid in the past too.

These various examples of how AI is being used to automate, manipulate and censor audio are quite rightly ringing alarm bells. Greater automation seems to be inevitable but it’s perfectly reasonable to question whether or not you want to live in a world where machines increasingly decide what’s in your best interests.

 

Facebook reportedly getting into the smart speaker game

All the other US tech giants have one, so it looks like Facebook is feeling left out and will launch some smart speakers later this year.

The news has been leaked by Digitimes, a Taiwan-based tech news site that specialises in tapping sources in the supply-chain channel to get clues about upcoming products. It reckons Facebook will launch two smart speakers with screens in the middle of this year, with the product strategy of making easier to interact with your Facebook friends, especially via video chat.

As you might expect once of them is said to be the basic model and the other the deluxe version, with all the bells and whistles. The basic one is codenamed Fiona and the better one is codenamed Aloha. Apparently Aloha will be marketed as Portal and will have clever gizmos like facial recognition, some extra social networking functions and some music licensing contracts.

If this report is accurate then it would seem to represent the latest manifestation of Facebook’s slow-motion panic attack in response to multiple competitive and existential threats. Not only is there growing evidence that Facebook users are using the service less than they used to, but there are growing concerns for it to take responsibility for all content published on Facebook, including acting as a censor.

Facebook resisted trying to get into the smartphone game, having done a good enough job with its app to render such futile gestures unnecessary. But it’s presumably worried that people will increasingly interact with the internet via smart speakers such as those offered by Amazon and, more recently, Apple.

Furthermore voice UI doesn’t really lend itself to Facebook, where the user experience is all about scrolling through posts and comments, and even less so to image-focused Instagram, which is owned by Facebook. So it’s easy to see why Facebook wants to get people using screens (other than TVs, tablets, smartphones, etc) in the living room.

But it’s hard to see how Facebook can possibly make a success of this. It’s very late to the market, has no track record in devices, and seems to be swimming against the current in trying to introduce a screen to devices defined by the voice UI. Also, because of the screen, these devices are likely to be relatively expensive, so what reason would anyone have to buy one of these instead of the alternatives?

Another report reveals Apple has had to drop its pants on margin just to get its smart speaker to market at a remotely competitive price point. And the reason the HomePod is so expensive is that Apple went all in on premium audio, but initial reviews indicate there is little to distinguish it from cheaper alternatives.

We would be happy to be proved wrong but this initiative, if it’s real, smacks of product development lead by sales and marketing rather than research and development. Products launched to defend a commercial position rather than as a genuine attempt to offer something useful usually fail, just ask Amazon, which basically wrote off its entire Fire Phone effort. Facebook is going to have to do something truly innovative to pull this off.

Intel is the latest to jump on the Amazon Alexa bandwagon

As Amazon continues to make the early running in the voice UI era with Alexa, Intel has created a special developer kit for it.

The Intel Speech Enabling Developer Kit is designed to enable developers to create consumer products featuring Alexa Voice Service. The reason a chip giant like Intel wants to be involved, other than merely jumping on the bandwagon, is that for voice UI to work well it not only needs decent processing but also a bunch of other sensors and distributed microphones.

“Natural language means machines need to clearly recognize and respond to user commands from a reasonable conversation distance,” said Miles Kingston, GM of Intel’s Smart Home Group, in a blog. “People speak and hear in 360 degrees, not just in a direct line of sight. Devices need array microphones and complex noise mitigation technology.

“A quality voice interaction means devices identify the speaker’s location, mitigate and suppress ambient noise, and understand spoken commands on the mics, even while playing music (talking and listening at the same time), as well as waking up when it hears the wake word (e.g. “Alexa”).”

Amazon seems to be doing a good job of partnering with other parts of the tech sector to boost its diversification efforts. Alexa is attracting both device and component makers, as well as retail partners, while AWS is showing a growing inclination to get into bed with strategic partners. In short Amazon is arguably the fastest growing of the internet giants right now.