What if AI doesn’t speak your language?

Episode Summary

Executive Summary: This Tech Life episode explores how AI is spreading into language access, public regulation, and aging support. It highlights Microsoft’s Lingua project to expand AI for underrepresented languages, contrasts European and US approaches to tech regulation, and reviews products like Sentai, Guardian, Nobi, and LEQ that aim to help older adults stay safe, independent, and connected.

Main Topics: AI for underrepresented languages (Priority: 5/5): Microsoft’s Lingua project seeks to create more training data and tools so AI systems can understand and serve languages with little online presence, helping close a potential AI digital divide. Language preservation and practical use (Priority: 4/5): The discussion connects language inclusion not just to access, but also to cultural preservation and real-world services such as translation, learning, and legal or agricultural support. Regulation of big tech in Europe vs the US (Priority: 5/5): Zach Myers explains that Europe emphasizes fundamental rights and market plurality, while the US is more permissive and innovation-led; AI is likely to sharpen these tensions. AI tools for older adults (Priority: 4/5): The episode reviews age-tech products designed to support older people through companionship, reminders, fraud protection, fall detection, and monitoring. Ethics and usability in age-tech (Priority: 4/5): Speakers stress that these tools must be supportive, accurate, and non-intrusive, especially when dealing with medication, fraud, loneliness, and care needs. Listener feedback on AI in public services (Priority: 2/5): The programme briefly revisits listener reactions to AI-assisted complaints to councils, showing public debate over whether AI is a nuisance or a response to unmet service obligations.

Key Arguments: AI access in a person’s own language is becoming as important as access to the internet itself; without it, people may be left behind in the digital economy. Projects like Lingua can help train models on languages with limited online content by partnering with local institutions, providing grants, compute, and expertise rather than centralizing all data collection. AI can support high-impact services in underserved communities, such as legal information in Malawi and agricultural advice in sub-Saharan Africa, but only if built with domain experts. Europe’s regulatory approach is rooted in fundamental rights and competition between multiple players, while the US is more comfortable with scale and market-driven innovation. AI regulation will likely diverge further across Europe, the US, and China, especially as AI agents begin making decisions and transactions on behalf of users. Age-tech products can genuinely improve safety and reduce isolation, but they must be designed carefully to avoid confusion, overreach, or false reassurance. Anti-scam tools and fall-detection systems show how AI can protect vulnerable users quietly in the background, but they rely on trust, permissions, and accurate detection. Companion robots and voice assistants may help older adults feel less lonely, but their value depends on ethical design and genuine usefulness rather than novelty. The public response to AI-written complaints suggests that technology is amplifying pre-existing tensions between citizens and institutions rather than creating them from scratch.

Data Points: Number of languages spoken worldwide: around 7,000 - Used to illustrate the scale of the language-access problem for AI systems. Country population: 22 million - Malawi’s population, mentioned in relation to legal-access challenges. Number of lawyers in Malawi: around 800 - Used to show why AI tools could help scale legal information access. Age-tech company competition size: 400 people / more than 100 projects - Guardian Labs’ founder said Guardian won first prize in a German AI competition. Top prize result: first prize - Guardian app won first prize in the competition described by Rita Burbosa. Podcast episodes: Season 4, all episodes available now - Promotional mention for Cyber Hack: The Conti Files.

Pivotal Quotes: "If you don't have access to AI in your own language, you might fall behind." — Juan La Vista Feres: Explaining the risk of an AI digital divide. "Europe has always been much more focused on fundamental rights." — Zach Myers: Describing the philosophical split between European and American tech regulation. "Beauty is a moral duty or that ugliness kills." — Roland Pilgrim: Explaining why Nobi’s safety technology is embedded in a ceiling light designed to look attractive.

Implications: AI is increasingly a language, policy, and care issue: if systems don’t support diverse languages or protect vulnerable users, inequality grows. Regulators and companies will face rising pressure to balance innovation with rights, safety, and inclusion.