Augmented Translators Driving Tech-in-the-Loop
Key Insights from the AI & Language Conference, September 12th 2024, Copenhagen
Last week’s language conference in Copenhagen shed light on how AI is transforming the audiovisual translation industry. It was packed with interesting talks and insightful speakers, but I can narrow down my 3 top takeaways:
- Post-editing is not always faster than translation according to recent study
- Localization in the AI age calls for a tech-in-the-loop approach
- Use of AI tools involving data pulls from the internet raises copyright concerns
Content Workflows Based on Topic and Languages
AI, especially generative models, is enabling faster, more accurate, and context-sensitive translations across multiple languages. The translator is often replaced by a post-editor to increase productivity and cost-efficiency.
Consequently, some LSPs have calculated that translation turn-around-time can be reduced by 30 %, but is it true? Not entirely, according to the first large-scale investigation of translation and revision speed in human translation and post-editing of neural machine translation, based on real-world data.
According to the study Is post-editing really faster than human translation? by Silvia Terribile at University of Manchester post-editing is e.g. usually but not always faster than human translation and translation speed is highly variable.
AI serves as a powerful tool, but we need a more nuanced approach in terms of analyzing the content and language combinations before starting the work flow and we should avoid making generalizations across projects on the human effort.
A big thanks to Amalie Foss, Søren Kuhn, all the volunteers and speakers for a super relevant and interesting conference!
AI & Language Conference was supported by:
Dansk Journalistforbund – Medier & Kommunikation, Creative Europe MEDIA Desk Denmark, Nordisk Film & TV Fond, Subline, Svensk Medietext, Norsk audiovisuell oversetterforening, Forum for Billedmedieoversættere, AudioVisual Translators Europe
Custom Tools with Humans in The Driver’s Seat
Professionals can leverage AI to improve efficiency, but human oversight remains essential to ensure cultural and contextual accuracy. This was confirmed by Peter Bjerre Rosa from the Danish Public Servicebroadcaster, DR, who hired “a robot” to test a hybrid approach to live subtitling with the use of a custom GPT plus supporting speech recognition tools.
DR concluded that it is often faster to cherry pick tools and approaches for individual projects and create custom solutions that meet the requirements and expected output.
Striking a balance between automation and human expertise is key but we need to put the human at the center not the technology – a tech-in-the-loop approach where the augmented translator analyzes and creates custom solutions involving AI and human guided language technology.
Copyright Concerns Calls for More Legislation
Another important aspect to consider is the copyright now being challenged by AI. This calls for more legislation on the use of content generated in tools extracting copyrighted material from the internet. How much should we change to avoid copyright violations, and how can data owners avoid that copyrighted material from being used without consent? The current text and data mining rules in the Directive on Copyright in the Digital Single Market and EU AI Act are currently not sufficient in terms of protecting the data owners and creatives who publish texts on the internet.
Author:
Lajka Birkedal Hollesen
CEO, Localised Web
What Generative AI Means for SEO and Content Teams
The way people find content online is evolving rapidly. With the rise of generative AI tools like ChatGPT, Gemini, and Perplexity, content discovery is no longer limited to search engine results pages. These tools, known as Large Language Models (LLMs), process massive volumes of information and deliver synthesized, conversational responses that often bypass traditional links. For marketers, this shift introduces new challenges — and opportunities — in how content is created, structured, and distributed.
SEO Remains Essential, But It’s No Longer the Whole Picture
Traditional SEO is still highly relevant. Ranking well in search engine results, optimizing metadata, and creating authoritative, keyword-rich content are foundational practices that continue to drive visibility. But now, another layer of visibility is emerging — one that involves preparing content to be understood, extracted, and referenced by AI systems.
That’s where Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) come in. These frameworks extend traditional SEO principles into the realm of generative AI.
What Are GEO and AEO?
- Answer Engine Optimization (AEO) focuses on structuring content to directly answer user queries in formats favored by AI-powered search engines like Google’s SGE, Bing Copilot, or Perplexity. This involves clear formatting, concise language, and question-based headings that feed into answer boxes or featured snippets.
- Generative Engine Optimization (GEO) is about creating content that is easily understood and referenced by LLMs. It helps ensure that a brand’s insights or expertise are included in AI-generated answers, even when the content isn’t directly cited or linked.
In this emerging landscape, the goal isn’t just to appear on page one; it’s to be part of the answer. That means a shift in strategy, not a replacement of existing SEO efforts.
High-Authority Content Is More Valuable Than Ever
As LLMs determine which content to surface in their responses, they prioritize trust and authority. Content published in reputable media outlets, supported by data and written with clarity, has a greater chance of being included in AI-generated outputs.
Content marketing platforms such as PRNEWS.IO support this multi-layered strategy. They assist brands in placing content within reputable media outlets, thereby improving brand visibility across trusted sources. This not only strengthens traditional SEO signals like E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) but also increases the likelihood of being recognised by LLMs.
How to Future-Proof Your Content Strategy
To stay visible in this dual environment — search engines and generative engines — marketers should:
- Focus on user intent: Create content that addresses specific questions and needs.
- Use semantic, conversational keywords: Reflect how users naturally speak, especially when interacting with AI.
- Structure content clearly: Use headings, short paragraphs, and bullet points for easy parsing.
- Establish authority: Include reputable sources, data, and expert perspectives.
- Distribute strategically: Aim for placements in respected media outlets to increase reach and credibility.
- Combine SEO, GEO, and AEO: These strategies are complementary. A strong SEO foundation strengthens AI visibility, and vice versa.
The rise of LLMs doesn’t diminish the value of SEO; it elevates the importance of high-quality, well-structured, and widely distributed content. SEO, GEO, and AEO together represent a unified approach to building discoverability across both traditional search and emerging AI-powered platforms.
Brands that adapt their content strategies to consider both search algorithms and language models will be best positioned to stay visible, relevant, and trusted in this new landscape.
Author:
Alex Nigmatulin
CMO of PRNEWS.IO