Category: UNCATEGORIZED

The constant refrain in the tech world is that the world is about to be rocked by the combination of AI and robots. Therefore you’d think that startups that build either AI or robots would be onto a winner. However, there’s a serious misunderstanding going on.

What is becoming clear is that the closer companies are to actual robotics, the only way to compete will be in genuinely transformational hardware. The era where a robotic arm was considered innovative is long over.

Similarly, robotics as a sector won’t go anywhere without being married to powerful machine learning and visual systems.

Thus it is that startups that can do both AI and blend this with robotics, that might be either off the shelf robotic arms or tools, will position themselves far higher up the valuation stack.

So it’s signification that UK-based startup Karakuri has ‘opened the kimono’ on it’s plans to do just that.

Karakuri uses a combination of robotics, machine learning, optics, and sensors to deliver a robot which will make personalized, freshly prepared, high-quality meals. The advantage is that the robot can make something which matches exactly what the customer wants (no nuts and seed for instance, just this amount of dressing etc ) and the result can also minimize food waste.

The startup comes at the right time. Research shows that almost two-thirds of consumers globally now follow a diet that limits or prohibits the consumption of some foods or ingredients due to food intolerance, as well as following a specific weight loss diet.
Karakuri has now raised a £7 million seed investment, led by Ocado.

Karakuri’s technologies also allow restaurants to move away from mass pre-packaged meals and significantly reduce food waste.

The fundraise is includes investments from Hoxton Ventures, firstminute capital, and Taylor Brothers and will be used to further develop the company’s technology, strengthen its IP base and expand its team for global growth.

For Ocado, the investment means it can expand its value proposition in grocery, especially through Ocado Zoom, its new delivery arm.

Karakuri CEO and co-founder, Barney Wragg, says, “Consumer eating habits in and out of the home are changing rapidly as demand increases for healthier options that match specific dietary requirements. This growth in menu personalization is putting huge pressure on restaurants, cafes and other food retailers. These providers have historically relied on identically mass-produced meals to maintain their profit margins. By using robotics and machine learning, Karakuri’s systems provide localized micro-manufacturing within an existing restaurant, retail or commercial kitchen. Our systems prepare personalized meals onsite in real time to the exact requirements of each customer.”

Brent Hoberman, Karakuri’s founding chairman, co-founder of Founders Factory and General Partner at firstminute Capital says: “The time is now for robotics and AI to drive change in the restaurant and food services business. Barney and Simon have assembled a world-class team to go after one the next large markets to be enhanced by this technology. We are delighted that Ocado, a global leader in robotics and distribution, has chosen to invest in Karakuri to lead the innovation in this sector.”

Hoberman says startups like Karakuri are going to become more significant as we reach the tipping point where manual workers are becoming less and less available to do the kinds of work that used to be done in restaurants.

Karakuri emerged out of the Founders Factory incubator, but the back story to this startup is significant. It’s advisory board includes industry experts from ARM, Ocado, Imperial College, Bristol Robotics Lab and Edinburgh Centre for Robotic.

Bristol Robotics Lab, in particular, has generated a world-class reputation for it’s robotics accelerator.

Jumpin’ Jack Flash-sale, startup fans! If you love the idea of saving even more money on passes to the preeminent tech conference focused on early-stage startups (and hey, who wouldn’t?), this is your chance. We’re holding a 48-hour flash sale that shaves an extra $200 off the price of select passses to Disrupt San Francisco 2019.

The flash sale starts at the stroke of midnight (PT) on Thursday, May 9 and runs through 11:59 p.m. PT Friday, May 10. Imagine experiencing Disrupt SF for less than the super-early-bird price. Crikey, that’s a good deal (even if we do say so ourselves). Jump on this offer now and buy your pass before the clock runs out.

Disrupt events always feature the most fascinating, relevant speakers in tech and investment, and this year is no exception to the rule. Hear from experts like Chris Dixon, general partner at Andreessen Horowitz (a16z), who will join us for a fireside chat about — among other things — current societal views on cryptocurrency and what needs to happen for it to become commonplace. Learn from Rachel Haurwitz, employee number one at Caribou Biosciences, a company founded to commercialize the gene editing technology known as CRISPR. Plenty more extraordinary speakers will be announced in the coming weeks and months, so stay tuned!

Maybe you’re ready to launch your startup to the world by taking on other exceptional startups in our epic pitch competition, Startup Battlefield. Go head-to-head with startups hand-selected by TechCrunch editors. Receive extensive, free pitch coaching, VIP treatment at Disrupt and a shot at the $100,000 grand prize.

If you’re not quite ready to compete but want to showcase your startup for free in the Startup Alley expo hall, apply to be a TC Top Pick. TechCrunch editors curate this cadre of outstanding startups and those selected will win a free Startup Alley Exhibitor Package, a videotaped interview and more at Disrupt.

One application enters you for consideration for both TC Top Picks and Startup Battlefield. It’s easy to apply, it won’t cost a thing and either one could take your early-stage startup to the next level. Apply right here.

There’s so much more to experience at Disrupt — like world-class networking in Startup Alley and beyond. CrunchMatch makes connecting with the right folks even easier, by helping you curate and connect with just the people you need to meet to succeed. Plus, there are the Q&A Sessions, panel discussions, workshops, and the Disrupt SF Hackathon.

Connect with community and opportunity at Disrupt San Francisco 2019, which takes place on October 2-4. And if you take advantage of this 48-hour flash sale, you’ll save an extra $200 off select passes including Startup Alley Exhibitor Packages. Whatcha waitin’ for? Jump!

Interested in sponsoring or exhibiting at Disrupt SF? Fill out this form to get in touch.

Samsung has been understandably silent about the Galaxy Fold for the last couple of weeks. The company’s been reassessing issues with the foldable’s display after initially chalking problems with review units up to small sample sizes and user error. It’s tough to say how difficult and expensive a fix will be, but this surely isn’t the sort of press it was hoping for with its first to market device.

CEO DJ Koh is finally ready to talk about the Fold — or at least offer news that there will soon be news. The exec told The Korea Herald that Samsung, “has reviewed the defect caused from substances (that entered the device), and we will reach a conclusion in a couple of days (on the launch).”

What Koh appears to be referring to specifically are the gaps in the fold mechanism that allowed material to get behind the display, damaging it when pressure was applied to the touchscreen.

From the sound of things, Samsung is hoping to have an update on timing at some point this week or early next, at the latest. Koh added, “We will not be too late,” which the paper took to be a suggestion that the Fold will begin shipping earlier than expected.

Samsung no doubt is hoping to have it out sooner than later, but the Note debacle’s two recalls should serve as a reminder that these things ought not be rushed.

Edgewell Personal Care, which owns brands like Schick (razors), Banana Boat (sunscreen) and Wet Ones (moist wipes), is adding Harry’s to that list in a $1.37 billion acquisition.

Founded in 2013, Harry’s is part of the current wave of brands using the internet to sell products directly to consumers. (In addition to razors, it also sells shower and face care products, and operates the Flamingo brand of women’s razors.)

It’s a trend that the established consumer giants have noticed, with Unilever acquiring Dollar Shave Club and Procter & Gamble acquiring Walker & Company.

With the acquisition, Harry’s co-founders and co-CEOs Andy Katz-Mayfield and Jeff Raider will become co-presidents of U.S. operations for Edgewell. (Speaking of direct-to-consumer brands: Raider is also co-founder of Warby Parker.)

“The combination of Edgewell and Harry’s is a pivotal step forward in further transforming our organization and strengthening our competitive position and ability to drive sustained growth and value creation,” said Edgewell’s President and CEO Rod Little in a statement. “Building on Edgewell’s and Harry’s complementary strengths, our combined company will have leading brands and omni-channel capabilities that are essential to meet the needs of the modern consumer and win in today’s market environment.”

Harry’s had previously raised around $375 million in funding, according to Crunchbase. Edgewell says the payment will break down to roughly 79 percent cash and 21 percent stock, giving Harry’s shareholders an 11 percent stake in Edgewell.

The deal is expected to close by the end of the first quarter of 2020.

Twitter’s latest transparency report is out.

The social media giant said it received six percent fewer requests for user data from governments between June and December 2018 than during its previous reporting period.

According to the newly released data, Twitter received 6,904 government requests for information on 11,112 accounts — the exact number of requests as its most recent report but with fewer accounts affected. The company turned over some data in just over half of all cases.

The U.S. filed the most requests — a total of 2,092 requests for data on 3,860 accounts, representing about one-third of all global requests.

Twitter also had 30 requests for data on 108 Periscope video-streaming accounts, disclosing some information in 40 percent of cases.

The company said it had received permission to disclose two national security letters (NSLs), which are subpoenas issued by the FBI with no judicial oversight that often come with a gag order. Although the company didn’t publish the letters as it has done previously, the two affected users have been informed.

Twitter also said it suspended 166,513 accounts associated with the promotion of terrorism, down 19 percent on the previous reporting period, and suspended 458,989 accounts for violating its rules relating to child sexual exploitation.

Intello, the New York City-based Saas management platform, announced a $2.5 million extended seed round today, along with some product enhancements.

The round was led by Resolute Ventures . Harrison Metal and Magnetico Ventures also participated along with various individual angel investors including Zane Lackey from Signal Sciences, Chris Smoak from Atrium and Zach Sherman from Timber. Today’s investment brings the total raised to $4 million, according the company.

Mike Hirshland, a partner at lead investor Resolute Ventures, saw Intello helping customers deal with a serious and growing issue inside companies. “They are solving the pain of SaaS sprawl that every organization is facing by empowering modern IT, finance and security teams with better visibility,” he said in a statement.

When everyone can sign up for these services for free or with a credit card outside the purview of IT, this can lead to potential issues around security and compliance. Since last year’s $1.3 million seed round, the startup has expanded beyond simply understanding what SaaS products a company has to include a compliance component, says Barak Kaufman, co-founder and CEO at Intello.

“We are really focused on what we view as end-to-end SaaS management, which includes spend optimization and mapping out unused products and licenses. We are still still doing that, but as we are selling primarily to IT, and sometimes InfoSec directors, the product has evolved passed that to working on compliance,” Kaufman explained.

To help with that, Intello has updated the product to map redundant applications along with the ability to change, add or remove licenses for third-party applications. This gives IT more control over unsanctioned applications including removing them from the system or limiting access to them if that is the goal.

The product is also now integrated with popular single sign on tools, Okta and Onelogin, to help companies map the usage of all the SaaS tools being used, whether free or paid, through their SSO tools.

The company offers more than two dozen integrations out of the box so far including popular SaaS tools like Salesforce, Box, Zendesk, Google, Slack and Office 365. It plans to use some of today’s funds to build dozens more with a goal of having 100 integrations by the end of this year.

Intello currently has 12 employees and 900 companies using the free and paid versions of the solution. Customers include InVision, Sprinklr and Instacart.

Singapore has passed a controversial bill that could equip the government with extensive powers to police online media and free speech.

The bill was first drafted last month and, as had been expected, it passed 72-9 in Singapore’s parliament, dominated by the ruling People’s Action Party (PAP) party, late on Wednesday.

As we reported last month, the bill caused concern through its potential to stifle free speech since a key feature enables the government, and in factor any minister, to force “corrections” to be added to online content that is deemed to be “false.”

Beyond media, the flex also extends to social media. According to the law, those found to be “malicious actors” face a fine of up to SG$50,000 ($37,000) or five years in prison for their content. If posted using “an inauthentic online account or a bot,” the fine jumps to a maximum of SG$100,000 ($74,000) or a potential 10-year jail term. Platforms like Facebook and Twitter face fines of up to SG$1 million ($740,000) for their role in such situations.

Designed to cover “a false statement of fact… has been or is being communicated in Singapore” or cases where politicians believe that issuing a correction is “in the public interest,” the bill also claims reach overseas — or, rather, intended reach overseas. Politicians can trigger it in situations judged to be “in the interest of friendly relations of Singapore with other countries.”

It remains to be seen how much success the Singapore government will have with its efforts. Domestic media may well be under control — the World Press Freedom Index ranks Singapore 151 out of 183 countries and self-censorship is common — but influencing newsrooms based overseas and social networks will likely prove difficult.

Facebook, for example, last November resisted calls to remove content flagged as defamatory by the government. That clearly frustrated officials.

“This shows why we need legislation to protect us from deliberate online falsehoods,” the Ministry of Law wrote in an announcement at the time.

How a takedown would work — and how the government might access encrypted chats on apps like WhatsApp and Telegram, which are also part of its focus — also remains unclear at this point.

The law has been criticized by free speech groups.

“Singapore’s new ‘fake news’ law is a disaster for online expression by ordinary Singaporeans, and a hammer blow against the independence of many online news portals they rely on to get real news about their country beyond the ruling People’s Action Party political filter,” Human Rights Watch deputy Asia director Phil Robertson wrote on Twitter.

“Singapore’s leaders have crafted a law that will have a chilling effect on internet freedom throughout Southeast Asia, and likely start a new set of information wars as they try to impose their narrow version of “truth” on the wider world,” he added.

Human Rights Watch — which came out with strong criticism of the bill last month — was criticized by the Singapore government last month which hit back at “its long-standing practice of issuing biased and one-sided statements about Singapore.”

Meanwhile, on the more assistive end of the dissenting voices, the Asia Internet Coalition — a group that represents Facebook, Google, Twitter, LinkedIn, Line and others — penned an editorial in Singapore’s Straits Times newspaper suggesting changes to the bill.

The opinion piece — which, irony alert, is restricted by a paywall — recommended specific processes, an imperial body to vet decisions, exemptions for opinion articles, satire and more, as well as a request for “clear and well-defined language and scope.”

Robertson is concerned that other counties in Southeast Asia will take the ball Singapore has punted and run with it, thereby creating other restrictive online content policies. That’s already happened to some extent. In Vietnam, a draconian cybersecurity law went into operation on January 1 while Thailand passed a controversial law granting a wide scope of powers to authorities in February.

Hotels are convenient but rarely homey. Short term rentals through Airbnb or HomeAway are often comfy but can be a pain to book and check-in. Business travelers often have to pick the best of two lousy options. Mint House is summed up best by Tige Savage, Revolution Venture managing partner: “Mint House is the best of a hotel without the worst of a hotel and the best of an Airbnb without the worst of an Airbnb.”

The New York-based Mint House is today announcing a $15 million financing round led by Revolution Ventures with participation from other investors and hotel industry veterans. The influx of capital and industry connections should go a long way in allowing the company to expand its offering that caters to business travelers looking for apartment-style accommodations with the predictability and reliability found in top-tier hotels.

Mint House is entering a crowded market dominated by Airbnb and monstrous industry incumbents. Mint House founder and CEO, Will Lucas, explained to TechCrunch how the company stands apart from hotels and short-term rentals. He said the company strives to provide the business traveler with a comprehensive hotel experience.

The service works a lot like a modern hotel. Travelers book online and proceed with their trip. Once the traveler arrives in the area or lands at the airport, a geofence is tripped, triggering directions to enter the building and room. The traveler doesn’t have to find someone to give them the key; their phone unlocks the door to an apartment-style room. Lucas says this takes a lot of stress off the property owner as the traveler does not need to bother anyone in the building — tenets or management alike.

Right now Mint House is focusing on markets in the US besides the top markets of New York City, San Francisco, and Los Angeles. Mint House is available in Indianapolis, Denver, Nashville, Miami, and Detroit. Lucas said he feels there are opportunities in these markets and often there are even worse accommodations available than in the busiest cities. Still, the total offering is relatively small: there are 200 rooms in operation, and 200 are scheduled to open by summer 2019. Travelers can book a Mint House through their app, website, or on travel sites like Expedia, Booking.com or Airbnb.

Mint House leases buildings from property owners, and Lucas was adamant that it’s the company’s goal to be a preferred partner with the real estate industry. He says that so far properties are receptive to Mint House’s business plan, and to be a global brand, the real estate community company needs welcome Mint House. Each property, even though some might be residential or mixed-use, have a dedicated staff of cleaners and management employed by Mint House. Likewise, each city Mint House operates in has a dedicated staff in-market overseeing the local operation.
“We understand that we have two customers – the multifamily developers and the business traveler – and by prioritizing both we’ve been able to secure A+ properties and deliver a top-rated guest experience,” Lucas said.” We have also formed a team of investors and advisors that are perfectly suited to help us differentiate ourselves in this lucrative market. With expertise and relationships in growing cities, as well as how to build world-renowned hospitality brands and services, we are confident that we can continue to build and expand alongside landlords and hospitality partners.”
Each market offers different regulations that Mint House has to follow. Lucas says the company supports whatever regulation they need to follow. It’s clear he’s trying to keeping Mint House on the right side of regulation. And he has help. Mint House has attracted the attention of several industry veterans that know how to grow a hospitality company.

Several notable hospitality veterans also participated in the financing round including Tom Mangas, the former CEO of Starwood Hotels, Carl Sparks, the former CEO of Travelocity, Kerry Hatch, the former president of St. Regis Hotels, and Rob Stewart, the Executive Vice Chairman of JBG Smith. Philippe Bourguignon, Vice Chairman of Revolution Places, former CEO of Club Med and Euro Disney and former president of Accor Hotels, Asia Pacific will join the board.

The deep network of industry insiders should help Mint House carve out a space of their own in the competitive hospitality world. It’s a tough market. Mint House has to get buy-in from business travelers and property owners alike. It’s a two-front battle. But as a frequent traveler myself, Mint House’s value proposition is compelling. I don’t need a hotel lobby or lackluster gym; I want a bed, couch and a TV without the terrible hotel menu system.

“Fake news” — news content that either misleads people with half-truths, or outright lies — has become a permanent fixture of the internet. Now, as tech and media platforms continue to search for the best way to fight it, Factmata — a London startup backed by Biz Stone, Craig Newmark, Mark Cuban, Mark Pincus, and more to build a platform to detect when false information is shared online — is announcing a new investor and partnership that will see it expanding its scope.

The company is picking up an investment from eyeo, the company behind Adblock Plus, and as part of it, Factmata is taking on the running of Trusted News, the Chrome extension that eyeo launched last year to give a nudge to those browsing content on the web to indicate whether a story is legit or shit.

Dhruv Ghulati, the CEO of Factmata — who co-founded the company with Sebastian Riedel, and Andreas Vlachos (Riedel’s other fake-news-fighting startup, Bloomsbury AI, was acquired by Facebook last year) — said that the financial terms of the deal were not being disclosed. He added that “eyeo invested both cash and the asset” and that “it’s a significant amount that strategically helps us accelerate development.” He points out that Factmata has yet to raise money from any VCs.

Trusted News today — an example of how it looks is in the screenshot above — has “tens of thousands” of users, Ghulati said, and the aim will be to continue developing and taking those numbers to the next level, hundreds of thousands of users by changing up the product. The plan will be to build extensions for other browsers — “You can imagine a number of platforms across browsers (eg Brave) search engines (eg Mozilla), hosting companies (eg Cloudflare) could be interested but we haven’t engaged in discussions yet,” he said — as well as to expand what Trusted News itself provides.

“The goal… is to make it a lot more interactive where users can get involved in the process of rating articles,” he said. “We found that young people especially surprisingly really want to get involved in debating how an article is written with others and engaging in rating systems, rather than just being handed a rating to trust.”

Ghulati said that eyeo’s decision to hand off running Trusted News to Factmata was a case of horses for courses.

“They are giving it to us in return for a stake because we are the best placed and most focused natural language understanding company to make use of it, and progress it forward fast,” he said. “For Factmata, we partner with a company that has proven ability to generate large, engaged community growth.”

“Just as eyeo and Adblock Plus are protecting users from harmful, annoying ads, the partnership between Factmata and Trusted News gets us one step closer to a safer, more transparent internet. Content that is harmful gets flagged automatically, giving users more control over what kind of content they trust and want to read,” said Till Faida, CEO & Co-Founder, eyeo, in a statement.

Factmata has already started thinking about how it can put some of its own technology into the product, for example by adding in the eight detection algorithms that it has built (detailed in the screenshot above that include clickbait, hate speech, racism, etc.). Ghulati added that it will be swapping out the way that Trusted News looks up information. Up to now, it’s been using a tool from MetaCert to power the app, a database of information that’s used to provide a steer on bias.

“We will replace MetaCert and make the system work at the content level rather than a list lookup, using machine learning,” he said, also noting that Factmata plans to add other signals “beyond just if the content is politically hyperpartisan or hate speech, and more things like if it is opinionated, one-sided, and or could be deemed controversial. “We won’t deploy anything into the app until it reaches 90% accuracy,” Ghulati said. “Hopefully from there, humans get it more accurate, per a public testing set we will make available for all signals.”

Ghulati himself is a machine learning specialist and while we haven’t heard a lot from Factmata in the last year, part of that is likely because building a platform from scratch to detect a problem that seems to have endless tentacles (like the web itself) can be a challenge (just as Facebook, which is heavily resourced and still seems to let things slip through).

He said that the eight algorithms it’s built “work well” — which more specifically he said are rating at more than 90 percent accuracy on Factmata’s evaluation sets on US English language news articles. It’s been meanwhile refining the algorithms on short form content using YouTube video transcripts, Tweets, Blog posts, and a move into adding more languages, starting with French.

“The results are promising on the expanded types of content because we have been developing proprietary techniques to allow the models to generalise across domains,” he said.

Factmata has also been working with ad exchanges — as we noted back when Factmata first raised $1 million, this was one of the big frontiers it wanted to tackle, since ad networks are so often used to disseminate false information. It’s now completed case studies with 14 major ad exchanges, SSPs and DSPs and found that up to 4.92 percent of a sample of pages served in some ad exchanges contain high levels of hate speech or hyperpartisan language, “despite them thinking they were clean and them using a number of sophisticated tools with bigger teams than us.”

“This for us showed us there is a lot of this type of language out there that is being inadvertently funded by brands,” he noted.

It’s also been gathering more training data to help classify content, working with people who are “experts in the fields of hate speech or journalistic bias.” He said that Factmata has “proven our hypothesis of using ‘expert driven AI’ makes sense for classifying things that are inherently subjective.” But that is in conjunction with humans: using experts leads to inter-annotator agreement rates above 68 percent, whereas using non experts the agreement of what is or is not a claim or what is or is not bias is lower than 50 percent.

“The eyeo deal along with other commercial partnerships we’re working on are a sign: though the system is not 100 percent accurate yet, within a year of building and testing our tech is ready to start commercialisation,” Ghulati added.

“Fake news” — news content that either misleads people with half-truths, or outright lies — has become a permanent fixture of the internet. Now, as tech and media platforms continue to search for the best way to fight it, Factmata — a London startup backed by Biz Stone, Craig Newmark, Mark Cuban, Mark Pincus, and more to build a platform to detect when false information is shared online — is announcing a new investor and partnership that will see it expanding its scope.

The company is picking up an investment from eyeo, the company behind Adblock Plus, and as part of it, Factmata is taking on the running of Trusted News, the Chrome extension that eyeo launched last year to give a nudge to those browsing content on the web to indicate whether a story is legit or shit.

Dhruv Ghulati, the CEO of Factmata — who co-founded the company with Sebastian Riedel, and Andreas Vlachos (Riedel’s other fake-news-fighting startup, Bloomsbury AI, was acquired by Facebook last year) — said that the financial terms of the deal were not being disclosed. He added that “eyeo invested both cash and the asset” and that “it’s a significant amount that strategically helps us accelerate development.” He points out that Factmata has yet to raise money from any VCs.

Trusted News today — an example of how it looks is in the screenshot above — has “tens of thousands” of users, Ghulati said, and the aim will be to continue developing and taking those numbers to the next level, hundreds of thousands of users by changing up the product. The plan will be to build extensions for other browsers — “You can imagine a number of platforms across browsers (eg Brave) search engines (eg Mozilla), hosting companies (eg Cloudflare) could be interested but we haven’t engaged in discussions yet,” he said — as well as to expand what Trusted News itself provides.

“The goal… is to make it a lot more interactive where users can get involved in the process of rating articles,” he said. “We found that young people especially surprisingly really want to get involved in debating how an article is written with others and engaging in rating systems, rather than just being handed a rating to trust.”

Ghulati said that eyeo’s decision to hand off running Trusted News to Factmata was a case of horses for courses.

“They are giving it to us in return for a stake because we are the best placed and most focused natural language understanding company to make use of it, and progress it forward fast,” he said. “For Factmata, we partner with a company that has proven ability to generate large, engaged community growth.”

“Just as eyeo and Adblock Plus are protecting users from harmful, annoying ads, the partnership between Factmata and Trusted News gets us one step closer to a safer, more transparent internet. Content that is harmful gets flagged automatically, giving users more control over what kind of content they trust and want to read,” said Till Faida, CEO & Co-Founder, eyeo, in a statement.

Factmata has already started thinking about how it can put some of its own technology into the product, for example by adding in the eight detection algorithms that it has built (detailed in the screenshot above that include clickbait, hate speech, racism, etc.). Ghulati added that it will be swapping out the way that Trusted News looks up information. Up to now, it’s been using a tool from MetaCert to power the app, a database of information that’s used to provide a steer on bias.

“We will replace MetaCert and make the system work at the content level rather than a list lookup, using machine learning,” he said, also noting that Factmata plans to add other signals “beyond just if the content is politically hyperpartisan or hate speech, and more things like if it is opinionated, one-sided, and or could be deemed controversial. “We won’t deploy anything into the app until it reaches 90% accuracy,” Ghulati said. “Hopefully from there, humans get it more accurate, per a public testing set we will make available for all signals.”

Ghulati himself is a machine learning specialist and while we haven’t heard a lot from Factmata in the last year, part of that is likely because building a platform from scratch to detect a problem that seems to have endless tentacles (like the web itself) can be a challenge (just as Facebook, which is heavily resourced and still seems to let things slip through).

He said that the eight algorithms it’s built “work well” — which more specifically he said are rating at more than 90 percent accuracy on Factmata’s evaluation sets on US English language news articles. It’s been meanwhile refining the algorithms on short form content using YouTube video transcripts, Tweets, Blog posts, and a move into adding more languages, starting with French.

“The results are promising on the expanded types of content because we have been developing proprietary techniques to allow the models to generalise across domains,” he said.

Factmata has also been working with ad exchanges — as we noted back when Factmata first raised $1 million, this was one of the big frontiers it wanted to tackle, since ad networks are so often used to disseminate false information. It’s now completed case studies with 14 major ad exchanges, SSPs and DSPs and found that up to 4.92 percent of a sample of pages served in some ad exchanges contain high levels of hate speech or hyperpartisan language, “despite them thinking they were clean and them using a number of sophisticated tools with bigger teams than us.”

“This for us showed us there is a lot of this type of language out there that is being inadvertently funded by brands,” he noted.

It’s also been gathering more training data to help classify content, working with people who are “experts in the fields of hate speech or journalistic bias.” He said that Factmata has “proven our hypothesis of using ‘expert driven AI’ makes sense for classifying things that are inherently subjective.” But that is in conjunction with humans: using experts leads to inter-annotator agreement rates above 68 percent, whereas using non experts the agreement of what is or is not a claim or what is or is not bias is lower than 50 percent.

“The eyeo deal along with other commercial partnerships we’re working on are a sign: though the system is not 100 percent accurate yet, within a year of building and testing our tech is ready to start commercialisation,” Ghulati added.