The rating agency EthicsGrade, focused on artificial intelligence (AI) governance, conducts an assessment of what it considers good practices related to digital ethics. And, in this, Facebook almost failed the year, receiving a C grade. The survey, released this Wednesday (31/03) by Tech Crunch, also analyzed the “ethical standards” of Twitter, VK and TikTok, among other social networks.

According to EthicsGrade, Facebook’s AI governance measures have shown that the company has strong cooperation on “data regulation with the European Union and other industry bodies.” However, scandals such as Cambridge Analytica's manipulation of data on social media during the US elections and the Brexit referendum in the UK were negatively impacted by the assessment. The C– rating corresponds to a value of 50,3. This also applies to Instagram and WhatsApp, as they both operate under the same structure.

“We want to be able to answer how TikTok compares to Twitter or WeChat to WhatsApp,” explains Charles Radclyffe, founder and founder of EthicsGrade. “And what we’ve found is essentially that things like GDPR [General Data Protection Regulation, u[a set of privacy laws in force in the European Union] have done a lot of good in raising the bar on issues like privacy and data governance. But in other areas we cover, like ethical risk or corporate practices versus public policy, these are more technical issues.”

To finalize the score, EthicsGrade used five criteria: structure, public policies, “technical barriers to trustworthiness,” ethical risk, and data privacy. Among the networks analyzed, Twitter was the one that obtained the best ethical evaluation, with a grade of B, ahead of VK, Instagram and Facebook, all with a grade of C. Below the average were LinkedIn, with 47,5, the Chinese platforms Renren, with 44,5, and TikTok, with 41,9, and the German Xing, with 39,9.

Algorithm-based search

EthicsGrade uses models based on artificial intelligence to create a more accurate sampling of the pattern ESG, a rating system that evaluates how companies and investments impact the environment and society and how this impact can lead to business or investment risks. The company uses NLP (natural language processing) to automate its data analysis, which includes tracking controversial topics and public statements.

According to Radclyffe, the AI ​​algorithm in the assessment is conditioned by a human-centered analysis. The data is organized to judge “the extent to which organizations implement transparent and democratic values, ensure consented information and risk management protocols, and establish a positive environment for errors and improvements”. The company also checks what measures are being put in place to manage the consequences of the use of algorithms and whether these are in line with digital ethics, transparency and democratic values.

“Our goal is to reach out to owners and shareholders,” Radclyffe says. “If you look at any of these companies like, say, Twitter, 29 percent of it is owned by five organizations: Vanguard, Morgan Stanley, Blackrock, State Street and ClearBridge. If you look at the ownership structure of Facebook or Microsoft, it’s the same companies: Fidelity, Vanguard and BlackRock. So we want to convince owners and shareholders that the questions that journalists have been asking for years are pertinent and relevant to their portfolios, and that’s how we plan to make our impact.”

The executive also explains that the content of social media posts is not included in the assessment. “What we are concerned about is how they govern their technology and where we can find evidence of that,” he explains. “What we then do is write to each company with our rating and assessment of them. We make it very clear that the rating is based on publicly available data.”

Through which channels you reach those people, classic and out of the box. TechCrunch

Image: Brett Jordan/Unsplash/CC