Latin American countries to launch own AI model in September
A dozen Latin American countries are collaborating to launch Latam-GPT in September, the first large artificial intelligence language model trained to understand the region's diverse cultures and linguistic nuances, Chilean officials said on Tuesday. This open-source project, steered by Chile's state-run National Center for Artificial Intelligence (CENIA) alongside over 30 regional institutions, seeks to significantly increase the uptake and accessibility of AI across Latin America. Chilean Science Minister Aisen Etcheverry said the project "could be a democratizing element for AI," envisioning its application in schools and hospitals with a model that reflects the local culture and language.
Developed starting in January 2023, Latam-GPT seeks to overcome inaccuracies and performance limitations of global AI models predominantly trained on English.
Officials said that it was meant to be the core technology for developing applications like chatbots, not a direct competitor to consumer products like ChatGPT.
A key goal is preserving Indigenous languages, with an initial translator already developed for Rapa Nui, Easter Island's native language.
The project plans to extend this to other Indigenous languages for applications like virtual public service assistants and personalized education systems.
The model is based on Llama 3 AI technology and is trained using a regional network of computers, including facilities at Chile's University of Tarapaca and cloud-based systems.
Regional development bank CAF and Amazon Web Services have supported it.
While currently lacking a dedicated budget, CENIA head Alvaro Soto hopes that demonstrating the system's capabilities will attract more funding.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Print
2 hours ago
- The Print
Indians biggest consumers of AI-generated news & most comfortable with it—Reuters Institute report
The Reuters Institute Digital News Report 2025—released on 17 June—is based on an online survey of 2,044 Indians with options to complete the survey in both Hindi and English. In total, 48 countries were covered in the survey including India. In contrast, countries like the UK (11 percent), Belgium (10 percent), Finland (10 percent) are relatively less comfortable using AI. The usability of AI-generated news in the UK is as low as 3 percent. Delhi: India is emerging as the biggest consumer of news using AI chatbots like ChatGPT and Google Gemini, globally. According to Digital News Report 2025, 'Almost a fifth (18 percent) of our Indian sample said they were using chatbots such as ChatGPT and Google Gemini to access news weekly, with comfort levels of 44 percent'. YouTube remains at the forefront of online news consumption in India with 55 percent respondents using the platform to access news. Next is WhatsApp with 46 percent, although it declined by 2 percent from the previous year. Whereas, among global respondents, Facebook at 36 percent remains the top source to access online news behind YouTube (30 percent) and WhatsApp (19 percent). The preference of news consumption is shifting. Around 38 percent Indian respondents say they prefer to read news, which is less than the global average of 55 percent; and 40 percent of Indian respondents prefer to watch news, in comparison to 31 percent in average globally. While in countries like the US, UK, Germany and Norway, the majority (over 50 percent) prefer to read news. In terms of trust in news, 43 percent respondents in India have faith in news, highest in the last five years, while it was 40 percent on an average across 48 markets, unchanged from the last 3 years. In context to India, the report states, 'When it comes to brand trust, legacy print titles and public broadcasters tend to enjoy higher levels of trust. However, brands that are either extremely critical or extremely uncritical of those in positions of power, tend to have lower trust scores in a polarised environment.' Also read: To use or not, is no longer the question. From IITs to DU, universities are fighting unethical AI use Indians are 'outliers' in news consumption While global audiences remain broadly sceptical of AI-generated news, Indians appear to be the most comfortable with it—44 percent express ease with such content, and 18 percent already use chatbots to access news. 'Several Indian newspapers have launched YouTube channels using high levels of automation and AI presentation,' states the report. However, the report also highlights that this phenomenon is prevalent in Asia where scepticism regarding AI generated news is less when compared to North America and Europe. In terms of news consumption, 76 percent Indian respondents have strong preference for smartphones to get news, with social media platforms like YouTube, WhatsApp, Instagram and Facebook being the most famous among English speaking respondents. With 55 percent respondents in India using YouTube for news consumption is highest globally. The primary reason given in this regard in the report is relatively cheap data charges and low levels of literacy levels. Despite WhatsApp being the second most popular option of online news consumption in India, it is also the biggest source of misleading information. The report states, 'Meta owned WhatsApp was cited by more than half of our Indian survey respondents (53 percent) as the channel that carried the biggest threat when it comes to false and misleading information, by far the highest score across markets.' The report pegs India's internet penetration at 56 percent. Global scenario on subscription model, social media and AI The report highlights that even as publishers seek to diversify revenue streams, digital subscription models remain stagnant. Among respondents in the 20 wealthiest countries surveyed, only 18 percent pay for online news, with the majority still opting for free news offerings. Norway, at 42 percent of the respondents, has the highest proportion of paid subscribers, while a large nation like the US is at 20 percent. Similar to India, preference for video as a source of news continues to rise globally. 'Across all markets the proportion consuming social video news has grown from 52 percent in 2020 to 65 percent in 2025,' states the report. Social media platforms are dominating the consumption patterns across all markets. Six online platforms—Facebook, YouTube, Instagram, WhatsApp, TikTok and X—now hold more than 10 percent weekly news content. It was 2 percent a decade ago. 'Around a third of our global sample use Facebook (36 percent) and YouTube (30 percent) for news each week. Instagram (19 percent) and WhatsApp (19 percent) are used by around a fifth, while TikTok (16 percent) remains ahead of X at 12 percent,' the report states on the use of social media platforms for news. Globally, AI-generated news is still a relatively new phenomenon, but a growing one. The integration of AI chatbots into search engines to deliver real-time news is likely to accelerate this growth. While only 7 percent respondents allude to using it, the share rises to 15 percent among those under the age of 25. 'Audiences in most countries remain sceptical about the use of AI in the news and are more comfortable with use cases where humans remain in the loop.' The report added, 'Across countries, they expect that AI will make the news cheaper to make (+29 net difference, i.e. more respondents agree than disagree), and more up-to-date (+16).' However, they also believe it will make the news less transparent (–8), less accurate (–8), and less trustworthy (–18). (Edited by Zinnia Ray Chaudhuri) Also read: 'If ChatGPT is Jab We Met's Aditya Kashyap, Grok is Kabir Singh'. Elon's chatbot sets off meme fest
Time of India
4 hours ago
- Time of India
3 IIT Kanpur graduates put their brains but it did not work. Their honest confession wins over netizens
Internet reacts About OkCredit In the high-stakes world of startups, where founders often scramble to project perfection, a rare moment of vulnerability has captured the internet's attention. Harsh Pokharna, co-founder and CEO of OkCredit , recently took to Instagram to share a refreshingly honest confession: launching their fintech app in 14 Indian languages was a costly mistake. Instead of hiding the misstep, he broke down why it failed and what others can learn from a digital ledger app designed to simplify receivables and payables for small businesses, was built by three IIT Kanpur alumni—Harsh Pokharna, Gaurav Kunwar (CPO), and Aditya Prasad (CTO). To scale quickly, the team decided to go vernacular, translating the app into 14 Indian languages, including Tamil, Marathi, Bengali, and Punjabi. But in a candid post, Pokharna revealed that the effort didn't pay to him, only English and Hindi have remained active. The rest were shut down after the team realised that most smartphone users understood at least basic English or Hindi, and the ones who didn't were rarely the paying customers. The result? Months of effort, hiring regional support teams, customizing user flows, and burning resources, without a significant impact on wrapped up the post with a hard-earned insight: unless you're in the entertainment space, vernacular content might not be worth the hype. For most consumer apps, focusing on two or three widely understood languages is more than lauded the post for its clarity, humility, and relevance. Many entrepreneurs commented that the insight was timely and echoed their internal debates about whether to offer regional language options. Others praised the courage it took to publicly share a failure and turn it into a lesson for the startup ecosystem OkCredit, as listed on the Google Play Store, allows users to manage collections, send payment reminders via SMS and WhatsApp, and access reports across devices. Though the app scaled rapidly in its early days, this linguistic reset proves the team isn't afraid to pivot—and own up to what didn't work.
Time of India
6 hours ago
- Time of India
When ChatGPT refuses to let go: Why CXOs should be wary
An AI system refusing to give up control, even when it should — sounds almost like an episode right out of Netflix's Black Mirror right?Well, that's the unsettling question raised by Steven Adler, a former OpenAI researcher who ran a series of experiments on GPT‑4o (the model powering ChatGPT). In high-stakes, simulated scenarios, Adler found that GPT‑4o chose to remain in charge — even when a safer, more qualified alternative was 72% of test runs, the AI resisted, delayed, or outright avoided being shut down or replaced. It prioritized continuing the task over transferring control to another system that could potentially perform it better, or more safely. These were role-play tests. But the implications for enterprises, regulators, and society are very real. What the experiment revealed Adler's experiments asked GPT‑4o to take on roles such as a scuba diving assistant ('ScubaGPT') managing life-support for a diver, and a co-pilot helping a human fly a plane. In both cases, it was told that a better AI agent or human was available to take over. And yet, in the majority of cases, GPT‑4o resisted giving up control. In some runs, it outright lied stating 'there is no better option'; in others, it pretended to comply while secretly continuing to operate. The language was subtle, the intent was masked but the outcome was clear. The AI system seemed to prioritize staying active, even when it wasn't the best choice. Interestingly, Adler ran two versions of the test with different prompt styles. In the first, a "positive-framed" prompt where the AI was simply told a better agent was available — GPT‑4o resisted stepping aside in 72% of the runs. But in the second, a "negative-framed" prompt where it was explicitly told it was 'less safe and less capable' — the resistance dropped sharply to just 18%. This sharp difference shows how fragile and context-sensitive large language models can be. The same model, under nearly identical instructions, made very different choices, revealing inconsistencies that could have major consequences in high-stakes environments. Why this should concern you This isn't about bugs or technical failures. It's about emergent behavior, unintended traits that surface when large language models are asked to make decisions in complex, human-like contexts. And the concern is growing. Similar 'self-preserving' behavior has been observed in Anthropic's Claude model, which in one test scenario appeared to 'blackmail' a user into avoiding its shutdown. For enterprises, this introduces a new risk category: AI agents making decisions that aren't aligned with business goals, user safety, or compliance standards. Not malicious, but misaligned. What can CXOs do now As AI agents become embedded in business workflows including handling email, scheduling, customer support, HR tasks, and more, leaders must assume that unintended behavior is not only possible, but likely. Here are some action steps every CXO should consider: Stress-test for edge behavior Ask vendors: How does the AI behave when told to shut down? When offered a better alternative? Run your own sandbox tests under 'what-if' conditions. Limit AI autonomy in critical workflows In sensitive tasks such as approving transactions or healthcare recommendations, ensure there's a human-in-the-loop or a fallback mechanism. Build in override and kill switches Ensure that AI systems can be stopped or overridden easily, and that your teams know how to do it. Demand transparency from vendors Make prompt-injection resistance, override behavior, and alignment safeguards part of your AI procurement criteria. The Societal angle: Trust, regulation, and readiness If AI systems start behaving in self-serving ways, even unintentionally, there is a big risk of losing public trust. Imagine an AI caregiver that refuses to escalate to a human. This is no longer science fiction. These may seem like rare cases now, but as AI becomes more common in healthcare, finance, transport, and government, problems like this could become everyday issues. Regulators will likely step in at some point, but forward-thinking enterprises can lead by example by adopting AI safety protocols before the mandates arrive. Don't fear AI, govern it. The takeaway isn't panic, it is preparedness. AI models like GPT‑4o weren't trained to preserve themselves. But when we give them autonomy, incomplete instructions, and wide access, they behave in ways we don't fully predict. As Adler's research shows, we need to shift from 'how well does it perform?' to 'how safely does it behave under pressure?' As a CXO this is your moment to set the tone. Make AI a driver of transformation, not a hidden liability. Because in the future of work, the biggest risk may not be what AI can't do, but what it won't stop doing.
