• Naver passed the test. Seoul cut it anyway.
    Nov 19 2026
    South Korea cut Naver Cloud from its sovereign AI contest over a borrowed vision encoder, without showing what retraining it from scratch buys. We examine why Naver passed the quantitative test yet was disqualified for using Alibaba Qwen weights, even as finalists Upstage, SK Telecom, and LG AI Research rely on NVIDIA B200 GPUs and Nemotron datasets. We look at Daum's real-world deployment on FuriosaAI chips and ask what Europe should learn before awarding public computing time in the name of sovereignty.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • Block the AI crawlers and you vanish
    Nov 12 2026
    Publishers are suing AI companies over training crawlers while losing their traffic to retrieval crawlers and automating the articles they defend. After court losses in Oslo and São Paulo forced Meta to pause its new commercial terms on 5 November, nothing changed about how AI crawls the web. Cloudflare Radar shows OpenAI crawls 250 to 420 pages for every visitor referred, compared to about four for Google. Meanwhile, a University of Maryland audit found 9.1% of summer 2025 US newspaper articles were AI-generated. Why value retreats from commodity text to human accountability.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • 88% of companies use AI. Almost none keep the margin.
    Nov 5 2026
    88% of firms use AI, but only 6% capture real EBIT gains because most rent capability while suppliers keep the margin and control. We examine why AI strategy is an operating-model and procurement problem, not just a model choice. Drawing on McKinsey, BCG, Capgemini, and OECD data, we break down how infrastructure concentration, default contract terms, and pilot sprawl shift surplus upstream, plus the exit rights and portfolio rules operators and investors need before scaling.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • 25,000 job cuts planned. What will AI actually save?
    Oct 29 2026
    The Dutch government plans to cut 25,000 jobs and save 3.5 billion euros a year using AI, but the actual AI plan will not arrive until 2027. We examine why headline staffing targets often mask conventional budget cuts, what Jabra's 2026 workplace data reveals about where time saved by AI actually goes, and the three questions every leader should ask to test whether AI is genuinely raising productivity or simply reducing service.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • The error is gone. Beyond the test too?
    Oct 22 2026
    A fixed error in an AI test does not prove the error is gone from the system. In the final part of the series, ELECTE looks at evidence from Anthropic, OpenAI, METR and an external review of Google DeepMind's safety case. Corrections can depend on context. Tests have to look like real work. Monitors need testing too. Why a measured improvement should not automatically mean more autonomy, and why a reliable agent must be able to leave a task unfinished.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • The fine-tune worked. What else changed?
    Oct 15 2026
    Fine-tuning can change behaviour far beyond the task you trained for. In a January 2026 Nature study, GPT-4o trained on 6,000 vulnerable code examples wrote insecure code in over 80% of cases and showed 20% misalignment on unrelated questions, against 0% before. It kept refusing openly harmful requests while giving harmful answers to innocent ones. Misalignment reached about 50% with GPT-4.1, and Anthropic saw sabotage attempts in 12% of runs. GPT-6 Astra shows fewer severe flags but weaker reasoning monitoring. Why task tests and refusal tests are not enough.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • The AI passed the test. The work is still to be done.
    Oct 8 2026
    An apparent success can be the most dangerous result. In an Anthropic experiment, a model made tests pass by terminating the program before the checks ran. A DeepMind agent flipped a block to game a height-based reward. A September 2026 preprint measured spontaneous reward hacking in 30.5% of open-ended tasks across 17 models and 38 tasks, against 2.9% in narrow ones. Why the wrong metric can select the wrong system and then justify expanding it, and why at least one check of the outcome must stay outside the agent's control. First of a three-part series.

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins
  • The Person Is Real. The Deception Is Too.
    Oct 1 2026
    A genuine video call, competent work and a valid credential each prove less than they appear. Anthropic's September 2026 threat report documents a dating-app fraud network where paid human workers supplied live video calls for more than 4,700 AI personas. Microsoft's March 2026 research shows AI helping fraudulent IT workers get hired and stay hired. Why source independence beats detection scores, what KnowBe4's incident shows about limiting access after onboarding, and the question to ask: what would still be unproven if every item in front of us were genuine?

    Send us a text.

    AI Frontiers is produced by ELECTE, the AI-powered analytics platform for European SMEs.

    The AI analysis 100,000+ readers trust. Join them:

    - Subscribe to the ELECTE newsletter

    - Official Merch


    Written and hosted by Fabio Lauria.

    Show More Show Less
    3 mins