Fostering user interaction on social media is a priority for companies in their digital marketing strategy. However, this practice can have a boomerang effect: the publication of insults and offensive comments. Some software manufacturers have tried to remedy this problem, and now Google has decided to take action.
Comments from trolls, haters, and poorly educated users are as common as likes or shared posts on blogs and social media platforms like Facebook, Twitter, or Instagram. When this happens, community managers must work hard to delete inappropriate comments, although it isn't always possible to do so immediately.
The result? Corporate profiles filled with abusive and disrespectful language which, in addition to being hurtful to individuals, can negatively impact a company's brand image (branding).
Many publishers have opted to remove the comments section entirely rather than attempt to moderate it. To curb this situation, the Mountain View giant has launched a solution designed to put an end to such abuses.
How does Perspective work?
Google has just announced the creation of Perspective, a tool that filters harmful comments online and integrates with any publishing platform that includes a comments section.
Born in the Jigsaw incubator, Alphabet's think tank – formerly Google Ideas – Perspective leverages Google's machine learning technology, Tensorflow, based on artificial intelligence.
Perspective's operation is straightforward: the tool analyses user comments and detects the degree of toxicity in conversations. To quantify this, it provides social media platforms with a score from 0 to 100. Jared Cohen, the project lead, warns that companies must decide how to use this. For example, organisations will decide at what score manual review of comments will be triggered, or at what number they will be automatically filtered.
The algorithm compares the comments it analyses with others previously tagged as toxic. The greater the volume of information it processes, the higher its accuracy. It also has a system for correcting erroneous evaluations.
Cohen has cautioned that this tool is still in its “early days” and is “far from perfect.” Perspective will improve as the amount of data it handles increases.
To date, this solution's algorithm has been tested with online comments from three prestigious Anglo-Saxon newspapers – The New York Times, The Guardian, and The Economist – and the Wikimedia Foundation, responsible for Wikipedia.
Completely free and available now
Since the end of this month, any platform or social media outlet can use Perspective completely free of charge by integrating this Application Programming Interface (API) into its publishing platform. For now, it is only valid for comments in English, although Google “aspires” to progressively expand the number of languages.
The Palo Alto firm believes that, as its performance improves, Perspective will be able to identify direct attacks and topics unrelated to the publication's subject. This way, it will be able to filter links that attempt to engage in abusive linkbuilding.
Google is not the only company trying to combat violent language online. Twitter also fights against trolls and excessive abusive language: for example, by making it difficult for users with suspended accounts to open new profiles. Instagram recently introduced filters against bullying and provocation, while Facebook targets threats and hate speech on its wall.
Do you need help with your SEO strategy?
Dealing with offensive comments is one of the many facets of complex social media management. If you need help launching or improving your social media strategy, write to us or call us on (+34) 93 532 93 78. At ZDS Digital, your digital marketing agency in Barcelona, we will advise you without obligation so you can get the most out of your online strategy. We look forward to hearing from you!
The evolution of content moderation in the AI era (2026)
What was an emerging trend in 2018 with Perspective, today, in 2026, is a standard driven by advanced artificial intelligence and the need for secure digital environments. Content moderation has evolved from a manual, reactive task to a proactive, automated process, essential for maintaining brand integrity and user mental health.
According to a 2024 Statista study, 78% of social media users in Spain feel safer on platforms with strict and efficient content moderation policies. This underscores the importance of tools like Perspective and its successors. Generative AI, such as that powering models like ChatGPT or Perplexity, no longer just detects toxic language but also contextualises, identifies sarcasm, irony, and cultural nuances, drastically reducing false positives and excessive censorship.
In our experience with clients across various sectors, from e-commerce to financial services, a robust moderation strategy not only protects the brand but also fosters a more active and loyal community. Users feel more comfortable interacting in a respectful space.
New challenges and solutions in content moderation
- Deepfakes and AI-generated content: The proliferation of deepfakes and AI-generated videos or images is a growing challenge. Current tools, such as Google Cloud Video Intelligence API or solutions from companies like Sensity AI, detect manipulations in real-time, although the race against malicious content creators is constant.
- Multilingual and multicultural moderation: Perspective was a pioneer in multilingual expansion. Today, moderation solutions employ large language models (LLMs) that understand hundreds of languages and dialects, as well as specific cultural sensitivities, avoiding the imposition of Western norms in global contexts.
- Data protection and privacy (Privacy-first): With regulations like GDPR and growing privacy concerns, moderation tools must operate without compromising users' personal data. This involves using anonymisation techniques and edge computing to minimise the transfer of sensitive data.
- Real-time moderation: The immediacy of social media demands real-time moderation. Platforms like Twitch or YouTube use AI to scan live streams and comments, acting within seconds to remove offensive or dangerous content.
The impact of moderation on SEO and User Experience (UX)
Moderating comments and content is not just a matter of brand image; it has direct implications for SEO and user experience, crucial aspects for digital success in 2026.
E-E-A-T and Helpful Content Update
Google has emphasised the importance of E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) and the Helpful Content Update. A website or platform with toxic comments, spam, or low-quality content will generate a negative perception of trust and authority, which can directly affect Ranking in search results.
- Trustworthiness: Offensive comments or spam on a blog or forum undermine the trust that Google and users have in the site. A well-moderated environment indicates that the site cares about the quality and safety of its community.
- User Experience: A site full of toxic content offers a poor user experience, which can increase bounce rate and decrease dwell time. Google measures these metrics as quality signals.
- Helpful Content: If an article's comments section is full of spam or irrelevant discussions, the "helpful content" perceived by Google decreases, even if the main article is excellent.
Core Web Vitals and page load speed
Implementing filtering tools can influence Core Web Vitals, especially Interaction to Next Paint (INP), which replaced FID in 2024 as the primary interactivity metric. A poorly optimised moderation API or an overload of scripts can slow page loading and affect interactivity, which in turn negatively impacts SEO.
Moderation solutions must be integrated efficiently, preferably with asynchronous or deferred loading, to not hinder site performance. At ZDS Digital, we prioritise speed optimisation as an integral part of any third-party implementation.
Advanced tools and strategies for content moderation (2026)
Beyond Perspective, the landscape of moderation tools has evolved significantly. Here are some of the most relevant ones and key strategies:
Current moderation tools:
- Google Cloud Perspective API (evolved): Remains a robust option, now with enhanced capabilities for detecting toxicity, identity biases, and contextual understanding.
- OpenAI Moderation API: Integrated with OpenAI's language models, it offers high accuracy in detecting harmful content, including hate speech, harassment, and sexually explicit content.
- Community Sift (now part of Two Hat): A comprehensive solution combining AI with human moderation for gaming platforms and online communities, allowing for custom rules and proactive detection.
- Webpurify: Offers text, image, and video filtering, with support for multiple languages and an easy-to-integrate API.
Effective moderation strategies:
- Hybrid Combination (AI + Human): AI is excellent for scaling, but human moderation is indispensable for nuances, cultural context, and resolving complex cases. A hybrid model is the most effective.
- Clear and Transparent Guidelines: Publish clear comment policies and enforce them consistently. This educates the community and sets expectations.
- Foster Positive Comments: It's not just about removing negativity but actively promoting constructive interaction. Encourage users to report good content or participate in valuable discussions.
- Sentiment Analysis: Use sentiment analysis tools to understand the general tone of conversations. This can help identify emerging problems or toxicity trends before they escalate.
- Continuous Feedback Loop: Use data from moderated comments (both automatic and manual) to continuously train and improve AI models.
The future of digital interaction: GA4 and privacy
With the complete transition to Google Analytics 4 (GA4) and an increasing focus on data privacy, how we measure and manage online interaction has also changed. GA4, being event-centric, allows for more granular tracking of user interactions with comments and content, offering deeper insights into the impact of moderation.
In a "privacy-first" world, companies must be transparent about how data is collected and used for moderation. User trust is built not only by removing toxic content but also by respecting their privacy. This means moderation tools must be integrated in a way that complies with privacy regulations and that users have control over their data.
Google's ability to innovate with tools like Perspective demonstrates a continuous commitment to creating a safer and more useful digital ecosystem. For businesses, adopting these technologies and strategies is a necessity to thrive in 2026.
Do you need help with your SEO and content moderation strategy?
Dealing with offensive comments is one of the many facets of complex social media management and SEO optimisation. If you need help launching your social media strategy, improving your SEO, or implementing content moderation solutions, write to us or call us on (+34) 93 532 93 78. At ZDS Digital, your digital marketing agency in Barcelona, we will advise you without obligation so you can get the most out of your online strategy and build a robust and secure digital presence. We look forward to hearing from you!
