Year: 2019

There are a lot of scientific papers out there, and finding the right ones, or the right connections between them, can be extremely difficult. Semantic Scholar uses AI to understand and index journal articles, but until recently has been limited to a handful of topics. It has now expanded to cover practically every branch of science — and some 175 million papers.

I covered Semantic Scholar, a project of the Allen Institute for AI, when it first launched in 2016, at which time it had only indexed papers in computer science and neuroscience. The next year, it added biomedical papers covering a variety of sub-topics.

The problem they are attempting to solve is simply that there’s too much information for academics to parse. And while they may do their best to keep up with the literature, a key insight or relevant result may be hidden away in an obscure journal that only gets the vaguest reference in a citation or review.

“We created it because of information overload in science,” explained project head Doug Raymond in an interview. “The focus of the team was, how do we make science more discoverable?”

Semantic Scholar uses natural language processing to get the gist of a paper, understand what processes, chemicals, or results are described, and make that information easily searchable. Not only does it make finding literature relevant to a given topic earlier, but it can establish patterns and find connections that might not have been clear before.

For instance it may be possible using the platform to identify trends in authorship as far as gender and other demographic balance (work on this is underway), or find bad actors who systematically cite themselves. In other cases the trends may be more immediately relevant: the majority of patients with kidney diseases are female, but the majority of those used in studies are male.

That’s not to say the system is doing research by itself, but facts and trends can appear under this kind of analysis that might have remained dormant otherwise. Especially since the system now encompasses most scientific domains and can make those connections between them as well as within them.

Expanding from a handful of disciplines to practically all of them was not an easy process, though the challenges are not what you might guess.

“We found that most of our models generalize well to new domains of science,” said Raymond. “That said, there’s always room for improvement. Some domains have very different conventions in how they write abstracts or lay out tables.”

The language model they created, SciBERT (an evolution of BERT, a more general purpose NLP agent), has been tweaked to understand different types of notation and so on. But apparently it didn’t choke, as I would have, after learning on CS and moving to organic chemistry. The results are functional enough to package into something like Supp.ai.

Raymond said the biggest problem was the more prosaic challenge of improving the system’s infrastructure to support the increased volume of data.

“The hardest thing, I’d say, was moving to a data pipeline that’s real-time and instantaneous rather than batch processing them,” Raymond explained. “Once we got to this scale, with the number of papers and partners, we had to redo the pipeline to get things done in hours rather than days.”

More partners means working with major science publishing outfits like Elsevier and Nature, who with the threat of SciHub and pressure from academics to move towards open access models, feel the presence of both stick and carrot as far as working with new efforts like Semantic Scholar.

As it is, the system has ingested most of the open access literature out there, and also has the key information for papers behind paywalls — users just won’t be able to pull up the full document without paying. On the other side of the equation, a partnership with Unpaywall keeps links to open access papers up to date. Open access articles, the platform has happened to note, are rapidly increasing proportion of all articles: more than doubling, from something over 10 percent to just under 30 in the last decade.

Now that the expansion part is mostly complete, the Semantic Scholar team is working on a few new features: improved summaries of articles, domain-specific functions, and a feed view that could show, say, a cell biologist the latest and most relevant findings in their field without exposing them to the firehose of research constantly being published.

Semantic Scholar is free to use — you can find it here.

Nooruldeen Agha, has been thinking about what’s next for fashion retail for years.

The serial entrepreneur behind the Dubai-based online fashion retailer, Elabelz and marketing studio Elephant Nation had always wanted to redesign the shopping experience for how customers actually shopped in stores and online.

“If it was 1994 and we knew what technology is today and we want to reinvent this [shopping] experience… one thought was how we bought our whole life and how we go to the mall,” says Agha. 

Shopping is, for most people, a social activity. Friends go to the mall or department store together to try on clothes and ask each other for advice. Most online and offline shopping experiences are completely divorced from that, Agha said.

So he wanted to take the social aspects of Instagram and the subscription box and retail elements of StitchFix to create the new Los Angeles-based startup, Flip Fit. 

But to do it, Agha needed a push. His businesses in Dubai were successful, he says, and there was no need for him to pursue another new venture — especially one in America.

Then he met Jonathan Ellman at the Summit conference in Los Angeles.

“We met at a party.  At midnight,” Ellman says. “At 10 o clock the next morning we were sitting on a balcony talking to each other and  came to an understanding that Noor with his dynamics and understanding the industry… that he could not stay in Dubai.”

Ellman has a history as an investor and an operator. He was the founder of the scout program at GreatPoint Ventures and spent years at HoneyBook. And he knew immediately that Agha’s idea had legs.

For the next year, the two laid the foundation for the business. Noor had all of the connections already. Elabelz was pulling in $23 million in revenue off of the sale of 150,000 boxes of clothes — so the logistics and fulfillment and brand partnerships would be a breeze. The company has 200 brands that have already signed on as of today’s launch including: AG, JBrand, Hudson, Retrobrand, Boyish, MadeWorn, Junkfood, Mavi and Edwin.

FlipFit works by creating a social network based on friends and followers. The company isn’t borrowing from Facebook or Instagram, but instead is trying to build out its network from scratch. Users of the app are encouraged to take vote on selfies their friends take in different outfits. Each vote garners in-app cash that can be redeemed whenever someone purchases an item ($10 for each new voting user and $1 per vote).

As users vote on the styles they like, they can also add clothes to a virtual wardrobe. When they’re ready they can select a few styles from that closet to be shipped out to them to try on. If the user doesn’t like the clothes, then they just return it.

The mechanics aren’t that different from a number of other online retailers, but the difference is in the company’s decision to create an entirely new social graph.

Initially, Agha and Ellman are tapping influencers to hook in their target customers. Over the next 90 days roughly 500 influencers across social media will be encouraging their audiences to vote on different outfits using the FlipFit app. The influencers are getting $150 in store credit twice-a-month or getting paid sponsorships (depending on the size of their following). The outfits with the most votes are the ones that the influencers will keep… training their audiences on the mechanics of how to shop as they market the product.

Agha says that the user experience is most akin to TikTok or Snap,rather than Instagram. There’s a publicly available feed for those who want to use it or the feed can be made private and shared among friends. And the app is only available for children 13 and up.

On the business side, the company is keeping 33% of the cash from any item sold. It’s cut is higher, because FlipFit handles all of the back end logistics of shipping and returns, according to the co-founders. Every box the company ships includes the standard pre-printed return label.

“Returns are our default. While the rest of the industry is fighting this phenomena, we are leaning into it,” said Jonathan Ellman, co-founder and co-CEO of Flip. “Almost half of all fashion shoppers bracket their online purchases, buying several pieces to try on at home with the intention of returning what doesn’t fit or what doesn’t match what they saw online. We believe returns should be as easy as the purchase and by making the shopping process more efficient and effective, we’re keeping clothes out of landfills and in your closet.”

The company is, to date, backed by a $3.75 million seed round led by TLV Partners with participation from Lool Ventures.

“Flip is the evolution of social media and e-commerce — birthing the baby of Istagram and Amazon and creating the first physical product marketplace where your likes and actions impact the products you receive,” says Rona Segev, a general partner at TLV.

Paid memberships and subscriptions have become the popular business model these days in the media industry, with publishers believing quite rightly that a closer connection to readers leads to a deeper customer relationship and ultimately more sustainable revenue.

That’s certainly true at big media companies (hello Extra Crunch) and there are also a spate of startups like Substack and Pico building out models for smaller publishers. But what if you want to build your own stack with an open-source foundation?

That’s where Ghost comes in. The open-source CMS, which has been around for a couple of years now following a successful Kickstarter, announced this week that it is launching subscription features as part of Ghost 3.0. The features will be included in both the open-source, self-hosted version as well as the organization’s paid hosted SaaS product.

Ghost 3.0 allows publishers to connect the CMS directly to Stripe, which processes credit card transactions as well as Apple Pay out of the box. Interestingly — and something CEO John O’Nolan empathized with me — Ghost will not take an additional transaction fee (I believe Stripe’s standard fees still apply). That’s a serious difference with some of its competitors, who take 10% or more of a publisher’s revenue as part of their business model.

From Ghost’s own website, “Unlike VC funded startups which pop up and shut down every few years, with Ghost you own the platform you depend on.”

In addition to handling payments, the CMS connects that payments data to a more robust set of user registration flows, allowing a publisher to know who its customers are and who is paying. Ghost content can be designated for different audiences — free content for everyone, and paid content only for active paying members. Ghost also has integrations with a number of other services.

Ghost itself has continued to expand from its April 2013 Kickstarter, which netted the organization $300,000 in pre-sales. CEO John O’Nolan shared that the organization has had net revenue of more than $5 million since launch, and has a current ARR of $1.73 million through its hosted product. That revenue funds a team of 15, who are fully distributed and remote.

Raydiant, a startup promising to turn TVs into interactive digital signs, is making several announcements today — a new company name, a new CEO, new partners and $7 million in new funding.

Until today, the company was known as Mira, and it was founded by Tuan Ho (who previously founded internet TV company Philo), along with venture studio Atomic. The goal is to offer an alternative to existing digital signage solution which it says are often improvised, bespoke or expensive.

With Raydiant, customers just plug the company’s HDMI device into a TV or other screen and connect that device to the internet. Then they can edit and update the content from Raydiant’s online dashboard, and they get  access to a number of other digital signage applications.

Customers include Westin, Ramada, Harvard University and Wahlburgers, the restaurant chain owned by the Wahlberg brothers. in fact, Raydiant/Mira was featured in an episode of the Wahlburgers TV show, with Mark Wahlburg testing out the technology as way to create a virtual presence in Wahlburgers restaurants.

The new applications being announced today show the range of what customers can do with these screens — the partners include BlueJeans (videoconferencing), Soundtrack Your Brand (licensed music), SinglePlatform (digital menus) and PosterMyWall (drag-and-drop content editing).

Ho and Atomic have also brought on Bobby Marhamat, the former COO of Revel Systems, as Raydiant’s new CEO.

“We are thrilled to have so much energy around the company with everything from our new brand, the financial support from our investors and incredible partnerships with the industry’s top enterprise companies,” Marhamat said in a statement. “Raydiant is a thriving and growing company, and we are committed to providing businesses with first-class service to bring any screen to life and create an interactive experience within a brick & mortar location.”

Lastly, the company is announcing that it has raised $7 million in funding led by 8VC, with participation from Atomic, Bloomberg Beta, Lerer Hippeau, SV Angel and Transmedia Capital. It says the money will be used to develop new applications, sign more partners, grow the team and bring on more customers.

With two months of 2019 still to go, Indian tech startups are already having their best year as a record amount of capital flows into the local ecosystem in a major rebound since the darkened funding environment in 2016.

The unlisted tech startups in India have raised $11.3 billion this year, a substantial jump from last year’s $10.5 billion fundraise, research firm Tracxn told TechCrunch.

This year’s fundraise, the largest sum for the local ecosystem in any year, further moves the nation’s burgeoning startup space on a path of steady growth. Since 2016, when tech startups accumulated just $4.3 billion — down from $7.9 billion the year before — flow of capital has increased significantly in the ecosystem. In 2017, Indian startups raised $10.4 billion, according to Tracxn.

Startups with consumer-facing offerings including financial services have attracted most of the capital this year — about $8.2 billion, Tracxn said. Following that is retail startups that have bagged about $2.3 billion and those that offer enterprise services with $1.5 billion. (There is some overlap of startups whose offerings fall under more than one category.)

Investors’ growing appetite for equity in India’s startups shows that the local ecosystem is maturing, said Dev Khare, a partner at VC fund Lightspeed Venture Partners . In an interview with TechCrunch, Khare noted that in 2014 and 2015, startups were largely focused on building e-commerce solutions and replicating ideas that worked in Western markets.

“But today, they are tackling a wide-range of categories and opportunities, and building some solutions that have not been attempted in any other market,” he said. He attributes much of this change to the arrival of telecom operator Reliance Jio and some government efforts such as introduction of GST taxation system for businesses and introduction of UPI payments infrastructure.

Jio, a three-year-old telecom operator run by India’s richest man, Mukesh Ambani, has disrupted the market with incredibly low-cost mobile data. The low-cost data meant that overnight tens of millions of Indians were able to come online for the first time.

This, alongside a cash crunch created by New Delhi in late 2016, led to a sudden explosion in demand for content and services including mobile wallets that created a massive opportunity for local startups to innovate, Khare said.

Financial services firm Paytm, which has raised more than $2 billion to date, has more than 200 million registered users in India, while Google Pay has amassed over 67 million active customers in less than two years of its existence.

Additionally, Khare said more people than ever in India today are willing to work at a startup. Citing Lightspeed’s internal research, he said four to five years ago, fewer than 20% of employees of a startup had ever worked for a startup before. Earlier this year, that figure had ballooned to more than 80%, he said.

There are some other promising signals as well: Of the top 150 Indian startups that raised capital in the first half of this year, 17.3% of them were either led or co-led by women, Indian news outlet The Morning Context reported on Wednesday, citing data from research firm Venture Intelligence. This is a massive jump from last year, when just 10% of startups counted women as their founder or co-founder.

A trend that appears to continue from the last several years is concentration of funds in a smaller number of startups. So far, tech startups in India have participated in 872 financing rounds, compared to 924 last year, and 1141 in 2017.

But that number, as well as total fundraise amount could change substantially by the end of the year as many more startups prep to close new financing rounds. Zomato, Swiggy, and Paytm alone are expected to close rounds worth as much as $3 billion in the coming months.

The third annual review of the EU-US Privacy Shield data transfer mechanism has once again been nodded through by Europe’s executive.

This despite the EU parliament calling last year for the mechanism to be suspended.

The European Commission also issued US counterparts with a compliance deadline last December — saying the US must appoint a permanent ombudsperson to handle EU citizens’ complaints, as required by the arrangement, and do so by February.

This summer the US senate finally confirmed Keith Krach — under secretary of state for economic growth, energy, and the environment — in the ombudsperson role.

The Privacy Shield arrangement was struck between EU and US negotiators back in 2016 — as a rushed replacement for the prior Safe Harbor data transfer pact which in fall 2015 was struck down by Europe’s top court following a legal challenge after NSA whistleblower Edward Snowden revealed US government agencies were liberally helping themselves to digital data from Internet companies.

At heart is a fundamental legal clash between EU privacy rights and US national security priorities.

The intent for the Privacy Shield framework is to paper over those cracks by devising enough checks and balances that the Commission can claim it offers adequate protection for EU citizens personal data when taken to the US for processing, despite the lack of a commensurate, comprehensive data protection region. But critics have argued from the start that the mechanism is flawed.

Even so around 5,000 companies are now signed up to use Privacy Shield to certify transfers of personal data. So there would be major disruption to businesses were it to go the way of its predecessor — as has looked likely in recent years, since Donald Trump took office as US president.

The Commission remains a staunch defender of Privacy Shield, warts and all, preferring to support data-sharing business as usual than offer a pro-active defence of EU citizens’ privacy rights.

To date it has offered little in the way of objection about how the US has implemented Privacy Shield in these annual reviews, despite some glaring flaws and failures (for example the disgraced political data firm, Cambridge Analytica, was a signatory of the framework, even after the data misuse scandal blew up).

The Commission did lay down one deadline late last year, regarding the ongoing lack of a permanent ombudsperson. So it can now check that box.

It also notes approvingly today that the final two vacancies on the US’ Privacy and Civil Liberties Oversight Board have been filled, meaning it’s fully-staffed for the first time since 2016.

Commenting in a statement, commissioner for justice, consumers and gender equality, Věra Jourová, added: “With around 5,000 participating companies, the Privacy Shield has become a success story. The annual review is an important health check for its functioning. We will continue the digital diplomacy dialogue with our U.S. counterparts to make the Shield stronger, including when it comes to oversight, enforcement and, in a longer-term, to increase convergence of our systems.”

Its press release characterizes US enforcement action related to the Privacy Shield as having “improved” — citing the Federal Trade Commission taking enforcement action in a grand total of seven cases.

It also says vaguely that “an increasing number” of EU individuals are making use of their rights under the Privacy Shield, claiming the relevant redress mechanisms are “functioning well”. (Critics have long suggested the opposite.)

The Commission is recommending further improvements too though, including that the US expand compliance checks such as concerning false claims of participation in the framework.

So presumably there’s a bunch of entirely fake compliance claims going unchecked, as well as actual compliance going under-checked…

“The Commission also expects the Federal Trade Commission to further step up its investigations into compliance with substantive requirements of the Privacy Shield and provide the Commission and the EU data protection authorities with information on ongoing investigations,” the EC adds.

All these annual Commission reviews are just fiddling around the edges, though. The real substantive test for Privacy Shield which will determine its long term survival is looming on the horizon — from a judgement expected from Europe’s top court next year.

In July a hearing took place on a key case that’s been dubbed Schrems II. This is a legal challenge which initially targeted Facebook’s use of another EU data transfer mechanism but has been broadened to include a series of legal questions over Privacy Shield — now with the Court of Justice of the European Union.

There is also a separate litigation directly targeting Privacy Shield that was brought by a French digital rights group which argues it’s incompatible with EU law on account of US government mass surveillance practices.

The Commission’s PR notes the pending litigation — writing that this “may also have an impact on the Privacy Shield”. “A hearing took place in July 2019 in case C-311/18 (Schrems II) and, once the Court’s judgement is issued, the Commission will assess its consequences for the Privacy Shield,” it adds.

So, tl;dr, today’s third annual review doesn’t mean Privacy Shield is out of the legal woods.

Thinking about sending a cold pitch? Curious how to connect with investors? Looking for fundraising tips for your first venture?

Startup funding experts including Forward Partners managing partner Nic Brisbourne, Target Global partner Malin Holmberg and DocSend co-founder and chief executive officer Russ Heddleston will sit down together and dish out all their best pieces of advice on the Extra Crunch Stage at TechCrunch Disrupt Berlin Dec. 11th and 12th.

Heddleston is making the trip out from San Francisco, where he operates a venture-backed business, DocSend, a platform that facilities secure document sharing analytics. Heddleston writes frequently for TechCrunch as a contributor. Recently, he penned this piece on the first mover advantage in startup fundraising and this one on the best times of year to seek venture capital.

Holmberg is joining us from London where she invests in high-growth tech companies across Europe out of Target Global, a fund with €700 million in assets under management. Holmberg previously focused on tech innovation at Tele2 in Stockholm and Vodafone in London. Before joining Tele2, she had stints at Marakon Associates, ATKearney and Morgan Stanley.

Finally, Nic Brisbourne, another London-based venture capital firm, will shed light on his career as an investor and more at our event in December. Brisbourne founded Forward Partners, a backer of Big Health, Zopa and more, in 2013 and invests in startups from the idea stage to the seed stage. Prior to launching Forward Partners, Brisbourne was a partner at Draper Esprit.

Make sure to stop by this session for realistic advice from venture capitalists and founders themselves.

Tickets to Disrupt Berlin are available here.

Starling Bank, the U.K.-based challenger bank founded by banking veteran Anne Boden, has raised an additional £30 million in funding.

In what was likely already agreed follow-on funding, perhaps contingent on milestones being met, previous backer Merian Chrysalis led the round with an investment of £20 million. JTC, another of Starling’s existing investors, also participated, adding a further £10 million.

Starling says the new funding will support increased investment in its consumer and SME bank accounts, as well as its B2B banking services. The capital will also be used to accelerate expansion into Europe.

Launched in May 2017, Starling has raised £263 million to date. In February this year it disclosed a £75 million funding round, and in the same month was awarded £100 million from the Capabilities and Innovation Fund, which was set up by Royal Bank of Scotland to fulfill European state aid conditions after a bailout during the financial crisis. Starling is using the CIF award to build out its SME account.

Meanwhile, the challenger bank says it is very close to reaching 1 million accounts opened. The number at time of publication was 930,000 accounts, with the 1 million mark expected within a few weeks.

That isn’t quite on par with competitors such as Monzo, Revolut or N26, all of which have several million opened accounts and have been growing significantly faster — even if that hyper growth isn’t without problems. However, Starling has always talked up its average bank deposits number as higher than other upstarts, evidence that consumers are using new banking offerings in different ways and not always as their primary salary account.

Cue statement from Boden: “This latest investment of £20 million from Merian Chrysalis will support Starling’s rapid growth and help us reach one million customers and £1 billion on deposit within weeks. It will also help us accelerate our global expansion, starting in Europe, so that even more people can benefit from the Starling app.”

Platform power is a helluva a drug. Do a search on Google’s Play store in Europe and you’ll find the company’s own Gboard app has an age rating of PEGI 3 — aka the pan-European game information labelling system which signifies content is suitable for all age groups.

PEGI 3 means it may still contain a little cartoon violence. Say, for example, an emoji fist or middle finger.

Now do a search on Play for the rival Fleksy keyboard app and you’ll find it has a PEGI 12 age rating. This label signifies the rated content can contain slightly more graphic fantasy violence and mild bad language.

The discrepancy in labelling suggests there’s a material difference between Gboard and Fleksy — in terms of the content you might encounter. Yet both are pretty similar keyboard apps — with features like predictive emoji and baked in GIFs. Gboard also lets you create custom emoji. While Fleksy puts mini apps at your fingertips.

A more major difference is that Gboard is made by Play Store owner and platform controller, Google. Whereas Fleksy is an indie keyboard that since 2017 has been developed by ThingThing, a startup based out of Spain.

Fleksy’s keyboard didn’t used to carry a 12+ age rating — this is a new development. Not based on its content changing but based on Google enforcing its Play Store policies differently.

The Fleksy app, which has been on the Play Store for around eight years at this point — and per Play Store install stats has had more than 5M downloads to date — was PEGI 3 rating until earlier this month. But then Google stepped in and forced the team to up the rating to 12. Which means the Play Store description for Fleksy in Europe now rates it PEGI 12 and specifies it contains “Mild Swearing”.

The Play store’s system for age ratings requires developers to fill in a content ratings form, responding to a series of questions about their app’s content, in order to obtain a suggested rating.

Fleksy’s team have done so over the years — and come up with the PEGI 3 rating without issue. But this month they found they were being issued the questionnaire multiple times and then that their latest app update was blocked without explanation — meaning they had to reach out to Play Developer Support to ask what was going wrong.

After some email back and forth with support staff they were told that the app contained age inappropriate emoji content. Here’s what Google wrote:

During review, we found that the content rating is not accurate for your app… Content ratings are used to inform consumers, especially parents, of potentially objectionable content that exists within an app.

For example, we found that your app contains content (e.g. emoji) that is not appropriate for all ages. Please refer to the attached screenshot.

In the attached screenshot Google’s staff fingered the middle finger emoji as the reason for blocking the update:

“We never thought a simple emoji is meant to be 12+,” ThingThing CEO Olivier Plante tells us.

With their update rejected the team was forced to raise the rating of Fleksy to PEGI 12 — just to get their update unblocked so they could push out a round of bug fixes for the app.

That’s not the end of the saga, though. Google’s Play Store team is still not happy with the regional age rating for Fleksy — and wants to push the rating even higher — claiming, in a subsequent email, that “your app contains mature content (e.g. emoji) and should have higher rating”.

Now, to be crystal clear, Google’s own Gboard app also contains the middle finger emoji. We are 100% sure of this because we double-checked…

Emojis available on Google’s Gboard keyboard, including the ‘screw you’ middle finger. Photo credit: Romain Dillet/TechCrunch

This is not surprising. Pretty much any smartphone keyboard — native or add-on — would contain this symbol because it’s a totally standard emoji.

But when Plante pointed out to Google that the middle finger emoji can be found in both Fleksy’s and Gboard’s keyboards — and asked them to drop Fleksy’s rating back to PEGI 3 like Gboard — the Play team did not respond.

A PEGI 16 rating means the depiction of violence (or sexual activity) “reaches a stage that looks the same as would be expected in real life”, per official guidance on the labels, while the use of bad language can be “more extreme”, and content may include the use of tobacco, alcohol or illegal drugs.

And remember Google is objecting to “mature” emoji. So perhaps its app reviewers have been clutching at their pearls after finding other standard emojis which depict stuff like glasses of beer, martinis and wine…

Over on the US Play Store, meanwhile, the Fleksy app is rated “teen”.

While Gboard is — yup, you guessed it! — ‘E for Everyone’…

Plante says the double standard Google is imposing on its own app vs third party keyboards is infuriating, and he accuses the platform giant of anti-competitive behavior.

“We’re all-in for competition, it’s healthy… but incumbent players like Google playing it unfair, making their keyboard 3+ with identical emojis, is another showcase of abuse of power,” he tells TechCrunch.

A quick search of the Play Store for other third party keyboard apps unearths a mixture of ratings — most rated PEGI 3 (such as Microsoft-owned SwiftKey and Grammarly Keyboard); some PEGI 12 (such as Facemoji Emoji Keyboard which, per Play Store’s summary contains “violence”).

Only one that we could find among the top listed keyboard apps has a PEGI 16 rating.

This is an app called Classic Big Keyboard — whose listing specifies it contains “Strong Language” (and what keyboard might not, frankly!?). Though, judging by the Play store screenshots, it appears to be a fairly bog standard keyboard that simply offers adjustable key sizes. As well as, yes, standard emoji.

“It came as a surprise,” says Plante describing how the trouble with Play started. “At first, in the past weeks, we started to fill in the rating reviews and I got constant emails the rating form needed to be filled with no details as why we needed to revise it so often (6 times) and then this last week we got rejected for the same reason. This emoji was in our product since day 1 of its existence.”

Asked whether he can think of any trigger for Fleksy to come under scrutiny by Play store reviewers now, he says: “We don’t know why but for sure we’re progressing nicely in the penetration of our keyboard. We’re growing fast for sure but unsure this is the reason.”

“I suspect someone is doubling down on competitive keyboards over there as they lost quite some grip of their search business via the alternative browsers in Europe…. Perhaps there is a correlation?” he adds, referring to the European Commission’s antitrust decision against Google Android last year — when the tech giant was hit with a $5BN fine for various breaches of EU competition law. A fine which it’s appealing.

“I’ll continue to fight for a fair market and am glad that Europe is leading the way in this,” adds Plante.

Following the EU antitrust ruling against Android, which Google is legally compelled to comply with during any appeals process, it now displays choice screens to Android users in Europe — offering alternative search engines and browsers for download, alongside Google’s own dominate search  and browser (Chrome) apps.

However the company still retains plenty of levers it can pull and push to influence the presentation of content within its dominant Play Store — influencing how rival apps are perceived by Android users and so whether or not they choose to download them.

So requiring that a keyboard app rival gets badged with a much higher age rating than Google’s own keyboard app isn’t a good look to say the least.

We reached out to Google for an explanation about the discrepancy in age ratings between Fleksy and Gboard and will update this report with any further response. At first glance a spokesman agreed with us that the situation looks odd.

Law enforcement tech outfitter Axon has announced that it will include automated license plate recognition in its next generation of dash cams. But its independent ethics board has simultaneously released a report warning of the dire consequences should this technology be deployed irresponsibly.

Axon makes body and dash cams for law enforcement, the platform on which that footage is stored (Evidence.com), and some of the weapons officers use (Taser, the name by which the company was originally known). Fleet 3 is the new model of dash cam, and by recognizing plate numbers will come with the ability to, for example, run requested plates without an officer having to type them in while driving.

The idea of including some kind of image recognition in these products has naturally occurred to them, and indeed there are many situations where law enforcement where such a thing would be useful; Automated icense plate recognition, or ALPR, is no exception. But the ethical issues involved in this and other forms of image analysis (identifying warrant targets based on body cam footage for instance) are many and serious.

In an effort to earnestly engage with these issues and also to not appear evil and arbitrary (as otherwise it might), Axon last year set up an independent advisory board that would be told of Axon’s plans and ideas and weigh in on them in official reports. Today they issued their second, on the usage of ALPR.

Although I’ll summarize a few of its main findings below, the report actually makes for very interesting reading. The team begins by admitting that there is very little information on how police actually use ALPR data, which makes it difficult to say whether it’s a net positive or negative, or whether this or that benefit or risk is currently in play.

That said, the very fact that ALPR use is largely undocumented is evidence in itself of negligence on the part of authorities to understand and limit the potential uses of this technology.

“The unregulated use of ALPRs has exposed millions of people subject to surveillance by law enforcement, and the danger to our basic civil rights is only increasing as the technology is becoming more common,” said Barry Friedman, NYU law professor and member of the ethics board, in a press release. “It is incumbent on companies like Axon to ensure that ALPRs serve the communities who are subject to ALPR usage. This includes guardrails to ensure their use does not compromise civil liberties or worsen existing racial and socioeconomic disparities in the criminal justice system.”

You can see that the ethics board does not pull its punches. It makes a number of recommendations to Axon, and it should come as no surprise that transparency is at the head of them.

Law enforcement agencies should not acquire or use ALPRs without going through an open, transparent, democratic process, with adequate opportunity for genuinely representative public analysis, input, and objection.

Agencies should not deploy ALPRs without a clear use policy. That policy should be made public and should, at a minimum, address the concerns raised in this report.

Vendors, including Axon, should design ALPRs to facilitate transparency about their use, including by incorporating easy ways for agencies to share aggregate and de-identified data. Each agency then should share this data with the community it serves.

And let’s improve security too, please.

Interestingly the board also makes a suggestion on the part of conscientious objectors to the current draconian scheme of immigration enforcement: “Vendors, including Axon, must provide the option to turn off immigration-related alerts from the National Crime Information Center so that jurisdictions that choose not to participate in federal immigration enforcement can do so.”

There’s an aspect of state’s rights and plenty of other things wrapped up in that, but it’s a serious consideration these days. A system like this shouldn’t be a cat’s paw for the feds.

Axon, for its part, isn’t making any particularly specific promises, partly because the board’s recommendations reach beyond what it is capable of promising. But it did agree that the data collected by its systems will never be sold for commercial purposes. “We believe the data is owned by public safety agencies and the communities they serve, and should not be resold,” said Axon founder and CEO Rick Smith in the same press release.

I asked for Axon’s perspective on the numerous other suggestions made in the report. A company representative said that Axon appreciates the board’s “thoughtful guidance” and agrees with “their overall approach.” More specifically, the statement continued:

In the interest of transparency, both with our law enforcement customers and the communities they serve, we have announced this initiative approximately a year ahead of initial deployments of Axon Fleet 3. This time period will give us the opportunity to define best practices and a model framework for implementation through conversations with leading public safety and civil liberties groups and the Ethics Board. Prior to releasing the product, we will issue a specific and detailed outline of how we are implementing relevant safeguards including items such as data retention and ownership, and creating an ethical framework to help prevent misuse of the technology.

It’s good that this technology is being deployed amidst a discussion of these issues, but the ethics board isn’t The Board, and Axon (let alone its subordinate ethics team) can’t dictate public policy.

This technology is coming, and if the communities most impacted by it and things like it want to protect themselves, or if others want to ensure they are protected, the issues in the report should be carefully considered and brought up as a matter of policy with local governments. That’s where the recommended changes can really start to take root.

Axon Ethics Report 2 v2 by TechCrunch on Scribd