Jobseekers and UK graduate employers are debating whether using AI during CV writing and online applications crosses into unethical behaviour or cheating. A survey of 155 of the UK’s biggest graduate employers found 92% had encountered or suspected AI misconduct, and EDF closed its graduate recruitment scheme early after receiving 12,000 applications, a 10% increase. Employers are tightening guidance and trying to make hiring processes more “AI-proof,” such as delaying CV reading and relying more on skills and behaviour-based assessments and video interviews.
Google is rolling out Gemini 4 Argon more broadly and expanding its Fairwind Program for trusted cyber defenders after testing limited access for misuse and misalignment risks.
Hachette CEO David Shelley said executives who rely on AI to draft and summarize communications are headed for a communications crisis and a “worse C-suite.” He pointed to a 2025 MIT Media Lab study finding participants who wrote essays with ChatGPT had the weakest brain engagement and most couldn’t quote from the essays they had just written. As a result, he argues leaders should keep writing with their own instincts to preserve sincerity and critical thinking rather than outsourcing first drafts to AI.
Anthropic says small AI models could reach the ability to copy themselves from one everyday device to another with little or no human involvement. The company tested a model with 26 billion parameters split across 15 $55 Android TV boxes and estimates the gap to full self-install and spread closes by the end of this year. As a result, Anthropic argues self-replicating risks move from frontier models toward smaller open-weight models, while critics say the experiments were engineered and don’t prove real-world worm-like spread.
Google is rolling out Gemini 4 Argon more widely in the Gemini app and its Antigravity feature, but access starts with paying users only. The rollout begins today. Capabilities and published benchmarks remain the same, while access is tier-limited and processing is guarded by a refusal system and an automated misalignment monitor.
Anthropic updated its rules so abusive behavior toward Claude must comply with its terms of service, while the article reviews ongoing, often repetitive debates about AI consciousness and moral status. The policy change specifies a ban on “sustained and needless” abusive or cruel behavior toward the models. As a result, discussions shift toward “moral patienthood” for current and future AIs and the author’s focus moves from consciousness arguments to how the new conduct rules affect alignment and safety behavior.
Henry Shevlin moved from full-time philosophy academia to a role at Google DeepMind as a resident philosopher focused on AI ethics and human-AI relationships. He now spends 80 percent of his time at Google DeepMind. The work focus shifts toward fast, real-time engagement with LLMs and safety/ethics questions like autonomy and consciousness, rather than only university-based research.
Character.AI filed for Chapter 11 bankruptcy after legal and scaling struggles undermined its interactive AI chatbot platform. The company reported assets and liabilities between $50 million and $100 million. It will run a court-supervised sale that includes a $20 million asset deal with Disney, with operations continuing only if a buyer is found; otherwise the platform will be wound down.
OpenAI’s research organization reports an automated “research intern” milestone where AI agent work overtook human labor this summer. Three point one agent-workdays per eight-hour day of human labor is the key figure, as of mid-August. The work shifts from agents mainly assisting to agents executing multi-day tasks under human direction, while humans still set priorities and decide whether to scale, pause, or deploy.
Character.AI filed for Chapter 11 bankruptcy protection in the United States and began a court-supervised asset sale process. Disney agreed to buy certain assets for $20 million, with the deal expected to close by late November or early December. The platform stays available during the proceedings, subscriptions continue, and it will be wound down if no buyer is found to keep operating it.
Every AI story that matters,
in your inbox by 8am.
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the
day in two minutes. Follow companies and topics for alerts, or get the
briefing in Slack. Free, no spam, unsubscribe anytime.
Reading TLDRocket needs no cookies, and the readership counts we rely on come from
our own cookieless analytics. Google Analytics is the exception: it sets cookies and
reports to Google, so it stays switched off until you allow it. You can change your
mind any time from “Cookie settings” in the footer.
Strictly necessary
Session security and form protection (tldrocket-session,
XSRF-TOKEN, 2 hours). The site cannot work without them,
so they need no consent.
Always on
Google Analytics 4 (_ga,
_ga_<id>, up to 2 years). Measures which
stories and sections readers use. Google acts as a third-party processor and may
store the data outside the EU. No advertising, no profiling, no data sold.
