Frequently Asked Questions


Q1: TrustedOut in two sentences?

A1: TrustedOut makes Intelligence trustworthy.
With its AI-operated media profiling, TrustedOut brings Corpus Intelligence to feed analytics tools, so business managers can make strategic decisions on sources they do trust.


Q2: What exactly are you delivering to TrustedOut customers ? (listings, whitelists, ratings?)

A2: Our users create and maintain their Corpus for which we permanently update via API or a dump upon requests, lists of media and relative sources (feeds) and, optionally article abstracts. We do not provide any ratings or scorings.


Q3: How do you keep the database up to date?

A3: The database is AI Operated.
. AI: It cannot be human operated. AI offers unbiased quantitative and qualitative profiling and this overtime and at all times.
. Operated: If and when AI encounters a challenge, AI will ask a librarian for a decision. Those decisions will teach our AI engine in case a similar situation arises.


Q4: What makes TrustedOut the best solution to fight fake news?

A4: Fake/False news are just the tip of the iceberg of misinformation and manipulations. In fact, thinking fact checking and blaming bad people is a very dangerous oversimplification of the issue at hand. Every high-tech major evolution has a societal impact. 20 years ago, blogs made everyone a publisher on the internet. Social networks made everyone editorialists, sharing opinions in real time. And everyone influences everyone. Consciously, honestly. Or not. Universal Truth is now challenged, meaning, Individual Trust must now be defined. This is the reason why, no-one, no organization can dictate what one will trust. Google News or Apple News, even after buying Texture, can’t say “We offer 200 trusted sources”. They can only say “We offer 200 sources you may trust”. Only the end-user can define their own trust.
That’s why we’re named “TrustedOut”. Our users define their own trust out of our super profiled media database.


Q5: Why is it so important?

A5: The trust in sources of knowledge, what we call “the corpus”, makes or breaks your education, your opinions and the decisions you make. IF YOU DON’T TRUST THE CONTENT YOU USE, YOU CANNOT TRUST ANY USE OF THAT CONTENT. Like the art of cooking, this is a transformation business. You aggregate ingredients, here content also named knowledge corpus, you transform it with a recipe or here some magic analytics algorithm and you eat or here, you make decisions. Garbage in, Garbage out. Bad in, Bad out. In other words, the quality of the Inputs dictates the quality of the Outputs.


Q6: Doesn’t a media database sound out-dated?

A6: [humour] Yes, we do use AI :) Not to be hype, but because it allows for an industrial solution to the classification and definition of the taxinomy, which used to be done by human documentalists, and thus was biased by design. At the end of the day, a database is a natural repository. Ours is elastic, making it more hype and glamour. :) [/humour]
Your real question, I believe, is on “why media” and not articles and AI magic or authors and influencers scoring?
Both do not work. Articles might, one day, fact check the content but again, misinformation could be, for the most part, true from a fact checking standpoint but written with a context and perspective and a culture that will make it misinformation for you. True fact can be important but is not the issue we are focusing on. At the author level, it does not work either because an author can write under a different name, write differently depending on his/her personal context and simply not be an expert on everything. Journalism guidelines can provide quality but we don't want to rate humans, we need an industrial solution. Humans cannot be it.
We’re left with Media.
Let’s set aside the business angle of a media. A media is a brand. A logo. And with it comes values. Values that make the reputation, the reason why people… trust that brand, i.e. that media. Fox News will always write as Fox News because people who go there, either love it or hate it. As Bill Ziff use to say: Love and Hate are the only two reasons people read magazines. Through this path of thoughts, we’ve created TrustedOut. An AI-Operated media database. Database and Media and AI, not because we gotta be cool but because AI today makes the dream of media classification true: Unbiasedly Industrial and Always-up-to-date.


Q7: But what makes the database unbiased?

A7: The solution to today’s misinformation, to be universal, must be industrial. Industrial means a machine with processes independent from any human emotions or bias.
Trust is not about Truth, which is community and temporary dependent. This is why TrustedOut is AI-Operated. The content of our database CANNOT be edited by a human. ALL information going in is from either, a reference site, such as Wikipedia for example or other corporate data sites, an AI-driven classification within our taxonomy or a self-declared information, such as feeds for example. ALL information are publicly exposed, not editorialized in any way, but verified and journaled permanently to track any suspicious changes over time. In other words, if the truth is community and temporary based, TrustedOut must collect information from communities and verify it at all time. Only AI can deliver this.


Q8: How can TrustedOut be universal?

A8: TrustedOut aims at being Universal, both for its media classification and its geo coverage with language independent AI supervision by librarians. The AI engine learns all the time and only in case of conflict will ask a librarian for advice. Again, NO human enters anything in the database but a librarian teaches the rules to our AI engine.


Q9: How will you make money?

A9: We are in a subscription to a database model. Simple and solid. Clients pay per corpus. Some options such as adding countries or abstracts will come with a fee.


Q10: Is TrustedOut a French company?

A10: TrustedOut will be both in France and in the USA. We'll expand farther later on.


Q11: Is your business re-selling content or syndication?

A11: We are not in the business of reading or publishing. Our Terms of Services (ToS) forbid any of these. Article abstracts are exclusively to feed text analytics tools.


Q12: TrustedOut uses multiple sources and reference sites : how can we prevent these sources from being biased?

A12: To avoid bias in reference sources, TrustedOut operates on all entries: a/ a cross-checked validation over several reference sites and b/ a tracking of those entries over time. All AI-operated, shall something suspicious happen, a librarian will be alerted and make a decision that will teach the AI engine to solve the issue in case it faces similar situations in the future.


Q13: Why do you focus on profiling sources and Media brands to fix trust and not on articles or authors?

A13: Trust is based on a reputation. An article does not have a reputation. An author may have a reputation but is temporary and linked to a matter. A publisher brand definitely has a reputation and its values guarantee stability. Bottom line: Your Trust is based on the publisher brands you value.