Spectrum Labs raises $32M for AI-powered content moderation that monitors billions of conversations for toxicity daily – TechCrunch

Two years into the pandemic, online conversations are still for many of us the primary interactions we have every day, and we collectively have billions of them. But as many of us have discovered, not all of these experiences are perfectly clean and positive. Today, a startup called Spectrum Laboratories — which provides artificial intelligence technology to platform providers to detect and stop toxic trading in real time (specifically, 20 milliseconds or less) — announces $32 million in funding. It plans to use the money to continue investing in its technology to double down on its growing consumer business and forge ahead in a new area, providing services to businesses for internal conversations and with customers, providing not only a means to help detect when toxicity is creeping into exchanges, but to provide an audit trail of activity for monitoring and broader trust and safety initiatives.

“We aspire to be language leaders where civility matters,” CEO Justin Davis said in an interview.

The round is led by Intel Capital, along with Munich Re Ventures, Gaingels, OurCrowd, Harris Barton and previous backers Wing Venture Capital, Greycroft, Ridge Ventures, Super{set} and Global Founders Capital. Greycroft led Spectrum’s previous $10 million round in September 2020, and the company has now raised $46 million in total.

Davis, who co-founded the company with Josh Newman (the CTO), said Spectrum Labs doesn’t disclose a valuation, but the size of the company today is a testament to how it’s doing.

Spectrum Labs works with just over 20 major platforms today – these include social media companies Pinterest and The Meet Group, dating site Grindr, Jimmy Wales Fandom entertainment wiki, Riot Games and the platform -form of online learning Udemy – which in turn has millions of customers sending billions of messages to each other every day, either in open chat rooms or in more direct private conversations.

Its technology is based on natural language and works in real time on both text interactions and audio interactions.

Davis notes that his audio work is “read” as audio, not transcribed to text first, which gives Spectrum clients a significant jump in activity response and counteracts what Davis called “the nature of the voice of the Wild West”, due to how slow responses are typically for those not using Spectrum’s technology: a platform must wait for users to report questionable content, and then the platform must find that audio in the transcripts, then she can take action — a process that can take days.

This is all the more important as voice services – with the rise of not just podcasting, but also services like Clubhouse and Spaces on Twitter – are growing in popularity.

Whether text or audio, Spectrum scans these exchanges for toxic content spanning more than 40 behavior profiles it originally built in consultation with researchers and academics around the world and continues to refine as it ingests more data from the web. The profiles cover parameters such as harassment, hate speech, violent extremism, scams, grooming, illegal solicitation and doxxing. It currently supports scanning in nearly 40 languages, Davis tells me, adding that it could work with any language, though Davis tells me there’s no language limit.

“We can technically cover any language in weeks,” he said.

The most visible examples of online toxicity have been in the consumer sphere – where they have taken place in open and more private online forums, hate speech and bullying and other illegal activities, an area where Spectrum Labs will continue to work and invest in technology. to detect ever more complicated and sophisticated approaches from malicious actors. One of Spectrum Labs’ goals will be to work on ways to improve how customers themselves can also play a part in deciding what they do and don’t want to see, as well as controls. and tools for a platform’s trust and safety team. It’s a tricky area, and arguably one of the reasons the toxicity has gotten out of control is that traditionally platforms have wanted to take a free-speech approach and not meddle in content. , because the other side of the coin is that they can also be accused of censorship, a debate that is still being played out today.

“There is a natural tension between what the policy implements and what users want and are willing to accept,” Davis said. His company’s view is that a platform’s job “is to keep out the worst of the worst, but also to provide controls for consumers to make a selection about what they want to see at over time”.

Along with this, Spectrum plans to shift more towards business services.

Opportunity in business is interesting because it not only includes how people within a business converse with each other (which could largely take a form similar to the consumer-facing services that Spectrum Labs already provides ), but also how a company interfaces with the outside world in areas such as sales, customer service, and marketing, and then leveraging the information Spectrum Labs collects in its analytics to potentially alter how each operates later. of these areas.

Certainly, this is not a market segment that has been ignored. Spectrum’s competitors here will include another startup in the conversation monitoring space, Aware, which focuses exclusively on the enterprise. (Light, meanwhile, is another contender in the consumer space.)

And there will certainly be others. We noted in our last article on Spectrum Labs that the founders and founding team came from Krux, a marketing technology company acquired by Salesforce (where they worked before leaving to found Spectrum Labs). I wouldn’t be surprised to see Salesforce play a more self-serving role in this area in the future, not least because it’s developing a very broad set of tools to help companies run their business more efficiently, and not just limited to CRM; but also because Bret Taylor, who once founded another social network and was the CTO of Facebook, now helps run the show, and may well have a particularly enlightened grip on how communication forums can be used and abused. .

For now, to address both consumer and business issues, Intel is entering as a strategic investor in this round, Davis tells me. The plan will be to integrate technology from Spectrum Labs to work more closely with Intel’s chip designs, which will further increase the speed at which it works, and Intel can use this as a unique selling point with potential customers of the hardware. of Intel as they place a higher priority on trust and security issues themselves.

“We believe that Spectrum Labs’ natural language understanding technology has the potential to become the central platform that powers the trust initiatives of thousands of companies around the world,” said Mark Rostick, vice president and chief executive of Intel Capital, in a statement. “As digital trust and ethical operations emerge as a key factor in helping organizations differentiate themselves, we see a huge opportunity to build a technology layer of trust and security into business operations.”

Comments are closed.