The Answer Engine AI News · Sydney

The AI industry is once again debating whether artificial intelligence poses an existential threat to humanity. This recurring conversation, discussed on TechCrunch's Equity podcast, reflects an ongoing split in the tech sector between those who believe advanced AI could pose catastrophic risks and those who view such warnings as overblown or distracting from more immediate concerns. These debates typically surface when major AI developments occur or when industry leaders make public statements about AI safety. The discussion touches on whether these warnings are genuine concerns about future AI capabilities, strategic positioning by companies seeking regulatory frameworks that might benefit incumbents, or attempts to generate publicity. For business owners, this represents background noise in the broader AI conversation – the kind of theoretical debate that dominates tech media but has little bearing on the practical AI tools most businesses are actually using today, like chatbots, automation software, and content generation tools.

Former US President Barack Obama has called on Democrats to prioritise artificial intelligence regulation, saying the party needs to make AI a "central agenda" item with a concrete plan to address safety and economic concerns. Obama's comments reflect growing pressure on US politicians to act on AI governance as the technology rapidly advances. While he didn't detail specific policies, his focus was on the economic disruption AI could cause—particularly job displacement—and safety risks that remain poorly understood. The statement adds to a chorus of voices, from tech leaders to academics, pushing for regulatory frameworks before AI deployment outpaces society's ability to manage its consequences. With Republicans likely to take a lighter-touch approach to tech regulation, Obama's call signals Democrats may campaign on stronger AI oversight in coming elections. For now, the US lacks comprehensive federal AI legislation, leaving most governance to voluntary industry commitments and patchwork state laws.

Perplexity, the AI-powered search engine, is now using OpenAI's latest GPT-6 Astra model to handle critical operational tasks with minimal human oversight. The company has delegated substantial responsibility to Astra, including writing communications, making changes to software code, and monitoring production systems. What's particularly notable is that Perplexity's team checks in on these automated processes "much less frequently" than they did with earlier GPT models, suggesting a significant leap in reliability and trust. This represents a shift from AI as a tool that needs constant supervision to AI as an autonomous operator handling mission-critical infrastructure. For context, production systems are the live environments that serve actual users, so any mistakes can directly impact service quality. The fact that a major search company is comfortable letting an AI model make changes in this environment, with reduced human monitoring, indicates OpenAI has achieved a new level of model dependability. This deployment suggests GPT-6 Astra can maintain more consistent reasoning, make fewer errors, and handle complex, multi-step workflows without the guardrails previous models required.

OpenAI has published details on how it transformed Habitat, originally a Python library, into a massive distributed storage system that now handles ChatGPT's global infrastructure. The platform serves over 1 billion users and processes 22 million requests per second. This technical evolution was necessary to support ChatGPT's explosive growth from a research project to one of the world's most-used AI services. The engineering challenge involved building a system that could scale horizontally across multiple data centres worldwide while maintaining speed and reliability. OpenAI's approach focused on distributed architecture rather than relying on a single centralised storage solution. The blog post provides insight into the infrastructure decisions required to run a consumer AI product at unprecedented scale. For context, processing 22 million requests per second puts ChatGPT's backend among the largest distributed systems globally, comparable to major tech platforms. While the post is primarily technical documentation aimed at engineers, it reveals the massive infrastructure investment required to deliver AI chat services that feel instant to end users.

OpenAI has announced that Cognition is using GPT-6 Astra to improve Devin, its AI software engineering tool. The integration specifically enhances Devin's ability to test the code it writes and demonstrate that the software actually works as intended. The goal is to reduce the amount of time human engineers spend reviewing AI-generated code, allowing them to ship products faster. Devin is an AI agent designed to handle coding tasks autonomously, and adding GPT-6 Astra's capabilities means it can now better validate its own work before passing it to human reviewers. This development is part of a broader trend where AI tools are becoming more self-sufficient, moving beyond just writing code to also testing and verifying it. For businesses using AI coding assistants, this could mean fewer bugs slipping through and less time spent on quality assurance. The announcement signals OpenAI's push to make AI development tools more reliable and production-ready, potentially accelerating software development cycles across industries that rely on custom applications or frequent updates.

Google has acknowledged that its Search Console tool doesn't properly report how websites appear in AI-powered search results. Search Console is the free tool most businesses use to track their Google search performance – showing which queries bring visitors, how often their site appears, and where it ranks. But as Google rolls out AI Overviews (the AI-generated answers that now appear at the top of many search results), the data in Search Console isn't capturing the full picture. This means business owners and SEO professionals can't accurately see whether their content is being featured in these prominent AI answers, how often it's cited, or what search terms trigger those citations. Without this visibility, it's difficult to understand whether your SEO efforts are actually reaching people through AI search features, or to adjust your content strategy accordingly. Google hasn't announced a timeline for improving this reporting, leaving businesses in the dark about a growing portion of search traffic and visibility.

Pet food brand Freshpet has shared a practical case study on how they improved their visibility in AI-powered search results, working with agency Intero Digital. The playbook covers three key areas: how to audit where and how AI systems currently cite your brand, how to restructure existing content so AI assistants are more likely to reference it, and how to connect this work to measurable business outcomes rather than treating it as a speculative experiment. While the original article doesn't provide specifics on Freshpet's results, the framework addresses a challenge many businesses face: traditional SEO tactics don't always translate to appearing in ChatGPT, Perplexity, or Google's AI Overviews. The approach focuses on making content more "citation-worthy" from an AI's perspective—presumably through clearer structure, authoritative signals, and formats that language models prefer to reference. For Australian small businesses wondering whether GEO is worth the investment, the emphasis on tying efforts to business priorities (not just rankings) offers a more practical entry point than chasing every new AI platform.

French newspaper Le Monde reports that Google's AI Overviews feature — which displays AI-generated summaries at the top of search results — hasn't caused a drop in their overall audience since launching in France. This is significant because many publishers have worried that AI Overviews would keep readers on Google instead of clicking through to their websites. However, Le Monde notes the traffic picture is more complicated than simple visitor numbers suggest. They're seeing growth in app traffic and subscriber numbers, which means the total story about how AI search features affect publisher traffic isn't straightforward. It's worth noting this is just one publisher's experience in one market, and the longer-term effects remain unclear. The key takeaway is that early fears about AI Overviews decimating publisher traffic may not be playing out uniformly — at least not yet, and not for everyone.

Anthropic's chief executive has called on AI companies to slow down their development of new AI models, citing concerns about potential misuse. The request comes as the industry races to release increasingly powerful AI systems, with competitors regularly announcing new capabilities and model releases. Anthropic, which makes the Claude AI assistant, appears to be advocating for a more cautious approach to rolling out advanced AI technology. The CEO's comments highlight growing unease within parts of the AI industry about the pace of development outstripping safety measures and proper oversight. While specific misuse scenarios weren't detailed in the source, the call for restraint suggests concerns about harmful applications, security risks, or unintended consequences of rapidly advancing AI capabilities. This represents a notable shift in tone from an industry that has been characterised by aggressive competition and speed-to-market priorities. The statement adds to ongoing debates about AI regulation, responsible development practices, and whether voluntary industry slowdowns are effective without coordinated policy frameworks.