In partnership with
Anthropic’s October 12 report quantifies a growing frontier-AI security concern. A party can query a powerful model at scale and use its responses to train another system. Anthropic says safeguards reduced reasoning tokens extracted from Claude on its own service by a factor of 10 to 100. That is a reported 90 to 99 percent decline, not an independent industry-wide finding.
Turn AI into Your Income Engine
Ready to transform artificial intelligence from a buzzword into your personal revenue generator?
HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.
Inside you'll discover:
A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential
Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background
Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve
Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.
Model distillation is not inherently improper. It trains a smaller model using outputs from a more capable one, often to lower deployment costs. The security issue here is unauthorized extraction of another developer’s model. Anthropic says reasoning traces can be more useful than final answers because they expose intermediate problem-solving steps.
Providers also use distillation internally, so authorization and data handling distinguish routine optimization from extraction disputes.
Anthropic describes several defenses. Its preserved-thinking setting limits how newer API accounts can alter earlier context, which the company says attackers used to elicit reasoning. Access to reasoning transcripts is also restricted by model and organization. A two-stage classifier identifies extraction prompts; monitoring, tighter limits, and bans target repeated campaigns. Anthropic says internal and external red-teaming helped improve the classifier.
The metric has a narrow scope. Anthropic counts reasoning tokens extracted on its platform. It says the observed amount fell 10 to 100 times after its new classifier, alongside earlier safeguards. This does not establish that every attack failed. It does not show that a copied model lost a particular capability or that third-party traffic fell similarly. Anthropic says it has not yet faced a highly sophisticated, persistent attempt against these measures.
Take the prompts our creative team actually uses
The key to maintaining brand quality at scale? Mastering how to communicate with AI models and embedding them into your creative strategy.
Join our honest discussion with Ari Murray, Chief Digital Officer at Salt and Stone, and get 5 tips for expert LLM prompting with ready to run prompts.
These controls also affect ordinary users. Anthropic says some developers had to update integrations. Its classifiers can misidentify benign requests; the company estimates a false-positive rate below 0.1 percent. Since September 24, it says blocked requests are charged in selected high-confidence areas. This is a company-reported rate, not an independent audit of user experience.
OpenAI’s September 30 disclosure shows the issue is not unique to one provider. It reported a coordinated attempt to extract protected reasoning. A spike included 16,000 requests using a relevant pattern across more than 4,000 users. Related patterns appeared across a cluster exceeding 15,000 users. OpenAI says these counts cover attempted, not necessarily successful, extractions. It said no encryption was broken and that findings were shared with industry partners.
Attribution requires caution. OpenAI said a core cluster involved people associated with Moonshot AI. The Cloud Security Alliance notes that provider attribution claims have not been independently verified. Its research note also distinguishes legitimate distillation from unauthorized use of another provider’s outputs. These points separate a technique’s ordinary uses from the security allegations and do not independently confirm Anthropic’s measurement.
Anthropic and OpenAI warn that extracted capabilities could be reproduced without the source model’s safeguards, including in cybersecurity and biology. This is a stated risk, not a demonstrated outcome of every distillation process. The evidence does not show that all distillation is theft or inherently dangerous. Authorization, exposed data, downstream controls, and independent testing matter.
The disclosure points to an operational security race. Developers must watch for coordinated behavior, restrict sensitive reasoning, and adapt defenses while measuring false positives and integration burdens. Since traffic can pass through third-party clouds, platform controls matter alongside provider classifiers. Public evidence remains largely provider-reported. Independent assessment of methods, coverage, and cross-provider effects would help establish whether lower extraction on one platform translates into durable protection. This check-in is not a final verdict; continued testing will help readers assess future security claims.
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.
LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.
The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.
Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
