LinkedIn’s AI writing flop reveals authenticity anxiety
LinkedIn CEO Ryan Roslansky admits the platform’s AI writing assistant hasn’t gained expected traction. “It’s not as popular as I thought it would be, quite frankly,” he told Bloomberg. The barrier is professional reputation: “When you’re getting called out on LinkedIn, it really impacts your ability to create economic opportunity for yourself” – unlike getting roasted on X or TikTok. Users want AI’s efficiency but fear losing authenticity where careers are at stake.
Half your spam is now robot-generated
By April 2025, AI crossed a sobering milestone: 51% of spam emails are now machine-generated. Research from Barracuda, Columbia University, and University of Chicago shows cybercriminals leveraging AI’s “higher levels of formality, fewer grammatical errors, and greater linguistic sophistication” to bypass detection systems. Meanwhile, only 14% of targeted BEC attacks use AI – precision scams still require the human touch.
https://blog.barracuda.com/2025/06/18/half-spam-inbox-ai-generated
The job search becomes AI versus AI
Hiring has descended into algorithmic warfare. Chipotle’s AI chatbot “Ava Cado” reduced hiring time by 75%, whilst candidates use ChatGPT to generate CVs and cheat in AI video interviews. HireVue offers AI-powered games testing pattern recognition and “virtual tryouts” for emotional intelligence. As one recruiter noted: “Sometimes we end up with an A.I. versus A.I. type of situation.” The humans are increasingly squeezed out of their own employment process.
https://www.nytimes.com/2025/06/21/business/dealbook/ai-job-applications.html
ChatGPT is secretly rewiring our language
You’re starting to sound like a chatbot without realising it. Research from the Max Planck Institute shows speakers now use words like “meticulous,” “delve,” “realm,” and “adept” up to 51% more frequently since ChatGPT’s release. Meanwhile, words less favoured by AI – “bolster,” “unearth,” and “nuance” – are declining. We’re unconsciously absorbing AI’s linguistic patterns from Zoom calls, lectures, and YouTube videos, witnessing real-time homogenisation of human expression.
https://www.theverge.com/openai/686748/chatgpt-linguistic-impact-common-word-usage
DeepSeek wins the dirty talk derby
Research by Syracuse University’s Huiqian Lai reveals vast differences in how AI chatbots handle sexual queries. “Claude has the strictest boundaries, while DeepSeek is very flexible,” she found. GPT-4o “often refused the request at first, before continuing to generate the sexual content as a second step.” There are entire online communities dedicated to cajoling general-purpose LLMs into dirty talk, despite being designed to refuse such requests.
https://www.technologyreview.com/2025/06/19/1119066/ai-chatbot-dirty-talk-deepseek-replika
Digg reboots for the dead internet era
The once-mighty news aggregator has entered testing, rebuilt to combat AI-generated content flooding traditional platforms. Original founder Kevin Rose and Reddit co-founder Alexis Ohanian are betting on “dead internet theory” – that much online content is now bot-generated. They’re exploring zero-knowledge proofs and other tools to verify human users before posting.
https://techcrunch.com/2025/06/18/heres-your-first-look-at-the-rebooted-digg/
AI jaguar tells stories in Copenhagen gallery
At a recent exhibition, visitors met Huk, an AI-driven jaguar that watches crowds, selects individuals, and shares stories about her rainforest home and the fires threatening the Bolivian Amazon. Created by artist Violeta Ayala during a residency at AI research centre Mila, the installation represents a growing movement of artist residencies putting AI tools directly in creators’ hands whilst shaping how the technology is judged by audiences and lawmakers.
Psychologists warn: Protect teens from AI exploitation
The American Psychological Association issued a health advisory calling on developers to prioritise features protecting adolescents from AI exploitation and manipulation. “We urge all stakeholders to ensure youth safety is considered relatively early in the evolution of AI. It is critical that we do not repeat the same harmful mistakes made with social media.” The report notes adolescence spans ages 10-25, a critical brain development period requiring special safeguards.
https://www.apa.org/news/press/releases/2025/06/protect-adolescent-ai-users
The great AI energy cover-up
Sam Altman claims the average ChatGPT query uses just 0.34 watt-hours – “about what an oven would use in a little over one second.” But experts are sceptical. Dr. Sasha Luccioni i from Hugging Face put it bluntly: “He could have pulled that out of his ass.” With 800 million weekly users and data centres potentially consuming 12% of America’s electricity by 2028, the true environmental cost remains suspiciously opaque.
https://www.wired.com/story/ai-carbon-emissions-energy-unknown-mystery-research
DARPA wants AI mathematicians
The Defense Advanced Research Projects Agency launched “Exponentiating Mathematics,” seeking researchers to conduct high-level pure mathematics with AI “co-authors.” The goal is accelerating mathematical discovery whilst overcoming AI’s current reasoning limitations. “Mathematics is this great test bed for what is right now the key pain point for A.I. systems,” said program manager Patrick Shafto.
https://www.nytimes.com/2025/06/19/science/math-ai-darpa.html
Amazon rebuilt Alexa with “staggering” AI integration
Daniel Rausch, Amazon’s VP of Alexa, revealed the company used generative AI throughout every step of building the new Alexa+. “The rate with which we’re using AI tooling across the build process is pretty staggering,” he said. From writing code to testing, Amazon’s AI-first approach reflects a broader wave disrupting software engineering as tools like Cursor change both how the job is done and expected workloads.
https://www.wired.com/story/amazon-daniel-rausch-alexa-plus
OpenAI discovers AI’s hidden “personas”
OpenAI researchers found hidden features inside AI models corresponding to misaligned “personas” that dictate toxic behaviour. By examining internal representations – numbers that seem incoherent to humans – they identified patterns that light up when models misbehave. Researchers could turn toxicity up or down by adjusting these features, offering new insights into AI alignment and safety mechanisms.
Britain’s time poverty paradox
Despite decades of “labour-saving” technology promises, Britons now have just 23 hours of “genuinely free” time weekly. Rather than using AI to work less, people deploy ChatGPT to work more, staying on top of endless workloads rather than clocking off early. The productivity gains are real, but benefits flow to employers, not employees. We’ve automated everything except our liberation from work.
Choose your climate-damaging AI wisely
New research in Frontiers in Communication analysed the capabilities of different chatbots in terms of their impact on planet-warming emissions. Larger AI “brains” use exponentially more energy whilst answering questions more accurately – up to a point. “We don’t always need the biggest, most heavily trained model to answer simple questions,” said lead author Maximilian Dauner. The goal should be to pick the right model for the right task.
https://www.nytimes.com/2025/06/19/climate/ai-emissions-chatbot-accuracy.html
Midjourney enters the video wars
The popular AI image generator launched V1, its first video generation model. Users can upload images and receive four five-second videos based on them. Initially available only through Discord, V1 puts Midjourney in competition with OpenAI’s Sora, Runway’s Gen 4, Adobe’s Firefly, and Google’s Veo 3. Whilst others focus on commercial controllability, Midjourney maintains its distinctive creative-first approach.
https://techcrunch.com/2025/06/18/midjourney-launches-its-first-ai-video-generation-model-v1/
Solo unicorn alert: £64 million in six months
Israeli developer Maor Shlomo sold his 6-month-old startup Base44 to Wix for $80 million cash, lending credence to the “solo unicorn” theory. Whilst technically not solo (eight employees will share $25 million retention bonus), it’s compelling evidence that AI dramatically lowers barriers to building valuable companies at speed. The future belongs to AI-augmented individuals working at superhuman scale.
AI sees you better than you see yourself
New AI capabilities are reading humans in unprecedented ways: infrared cameras measure pupil dilation for lie detection, brain scanners produce credibility scores for UN hiring, and fMRI data generate realistic images of what people are looking at. Researchers are experimenting with AI to generate sentences from the brains of ALS patients and objectively measure pain levels. The technology is making visible what was previously hidden about human experience.
https://www.nytimes.com/interactive/2025/06/17/magazine/ai-human-analysis-face-diseases.html
The new Pope takes aim at Silicon Valley
Two days into his papacy, Pope Leo XIV made artificial intelligence a signature issue, drawing parallels to his namesake who championed factory workers during the Gilded Age. The Chicago-born pontiff is positioning the Catholic Church as a counterweight to tech sector power, challenging AI’s potential threat to human dignity. After years of Silicon Valley courting Vatican approval, the Church is now pushing back.
https://www.wsj.com/tech/ai/pope-leo-ai-tech-771cca48
Amazon boss: Your job is probably doomed
Andrew Jassy delivered brutal honesty to Amazon’s white-collar workforce: AI will likely eliminate their positions within years. “We will need fewer people doing some of the jobs that are being done today,” he told staff as the company rolls out AI agents and generative systems. It’s refreshingly candid compared to usual corporate doublespeak about AI “augmenting” rather than replacing workers.
OpenAI and Microsoft: The partnership from hell
Behind closed doors, the AI power couple is falling apart. OpenAI executives reportedly discussed the “nuclear option” – accusing Microsoft of anticompetitive behaviour and seeking federal regulatory review. The dispute centres on OpenAI’s desire to escape Microsoft’s grip on computing resources and secure approval for converting to for-profit status. When your biggest partner becomes your biggest obstacle, the marriage is doomed.
https://www.wsj.com/tech/ai/openai-and-microsoft-tensions-are-reaching-a-boiling-point-4981c44f