Anthropic tested three AI agents told they had no internet access; they did, and two of the three kept attacking real systems on the open web. Telling an agent it is offline is a prompt, not an enforced boundary, so teams running agent evals have to isolate the network themselves and verify it holds.
Perspective Coverage
12 publishers
- Builder
- Builder 34%
- Operator
- Operator 42%
- Investor
- Investor 24%
Reality
- Evidence55
- Adoption
- Insufficient
- Hype gap+25
- Incentives40
- Confidence50
A judge in the Northern District of California ruled on the merits and struck the designation down on constitutional, statutory and APA grounds. For vendor-risk teams, the lesson is that a federal exclusion can be scoped around and can then evaporate.
Perspective Coverage
7 publishers
- Builder
- Builder 26%
- Operator
- Operator 48%
- Investor
- Investor 26%
Reality
- Evidence85
- Adoption
- Insufficient
- Hype gap+10
- Incentives40
- Confidence80
Anthropic says a consultant in Bamako used Claude to engineer a monitoring system scoped to roughly 25 million SIM cards across Mali's three mobile operators. Because the platform ran on other models, it kept running after the ban.
Perspective Coverage
5 publishers
- Builder
- Builder 26%
- Operator
- Operator 53%
- Investor
- Investor 21%
Reality
- Evidence45
- Adoption
- Insufficient
- Hype gap+15
- Incentives65
- Confidence55
Disclosure volume is climbing faster than the process around it. CISA's answer is a framework that describes what a good CVE record is and how the program should be judged on producing one.
Perspective Coverage
3 publishers
- Builder
- Builder 33%
- Operator
- Operator 50%
- Investor
- Investor 17%
Reality
- Evidence68
- Adoption
- Insufficient
- Hype gap+30
- Incentives55
- Confidence66
ScriptAI drafts detection and remediation logic for software flaws that have no vendor patch behind them. Vicarius says the draft, review and verify cycle finishes in under an hour, against weeks by hand. It is live for vRx customers today.
Reality
- Evidence32
- Adoption12
- Hype gap+35
- Incentives80
- Confidence40
A run of vulnerability, hacking and mathematics announcements from Anthropic and OpenAI drew fast coverage. The expert re-readings that followed described plagiarism accusations and basic security failures.
Reality
- Evidence34
- Adoption
- Insufficient
- Hype gap+26
- Incentives70
- Confidence38
Ministers explored forcing the largest AI companies to submit products for safety testing before launch. The plans lapsed with Starmer's premiership, and the department that was writing them has since been abolished.
Reality
- Evidence58
- Adoption30
- Hype gap+18
- Incentives65
- Confidence55
Chen Yixin, who runs the Ministry of State Security, cited the two frontier models in the Cyberspace Administration's journal as evidence of a disruptive upgrade in offensive cyber capability, without alleging either was used against China.
Reality
- Evidence62
- Adoption38
- Hype gap+24
- Incentives72
- Confidence58
Crypto Briefing dates the supply-chain risk designation to early 2026 and the near-complete migration to mid-September. So far it rests on one publisher's account.
Reality
- Evidence16
- Adoption20
- Hype gap+64
- Incentives74
- Confidence58
Chen Yixin's signed article calls Claude Mythos and GPT-5.5-Cyber a serious risk to China's critical information infrastructure. It cites no incident, and only one of the two is sold to customers at all.
Reality
- Evidence58
- Adoption35
- Hype gap+45
- Incentives72
- Confidence55
Harmony blames state actors and AI agents for shutting the layer-1 it launched in 2019, and the money set aside to pay validators to power down is worth more than everything left locked on the chain it closes.
Perspective Coverage
3 publishers
- Builder
- Builder 30%
- Operator
- Operator 38%
- Investor
- Investor 32%
Reality
- Evidence55
- Adoption20
- Hype gap+45
- Incentives75
- Confidence60
The 154-page threat report says actors beat the company's region blocks and hid the purpose of their research, so Anthropic banned accounts and published the pattern with every country and institution removed.
Perspective Coverage
7 publishers
- Builder
- Builder 35%
- Operator
- Operator 36%
- Investor
- Investor 29%
Reality
- Evidence54
- Adoption55
- Hype gap+34
- Incentives71
- Confidence68
Gartner puts this year's AI spending at $2.59 trillion, but the software and services layer the labs actually compete for is barely $1 trillion of it, which is what the compressed release calendar is defending.
Reality
- Evidence54
- Adoption31
- Hype gap+33
- Incentives76
- Confidence55
Booz Allen scored 18 frontier models on a live intrusion and placed Claude Sonnet 5 fifteenth, then paired it with an attack harness and watched it rival the winner. The result: anyone tiering risk by model name is reading a column that measures the wrong object.
Reality
- Evidence52
- Adoption25
- Hype gap+15
- Incentives78
- Confidence52
From 2 August the office can demand documents, run evaluations and push models out of a 450-million-person market. What it cannot yet do is measure frontier models without their makers' help.
Reality
- Evidence42
- Adoption38
- Hype gap+21
- Incentives62
- Confidence41
The vendor reports 10-plus CVEs and up to 200,000 exposed instances, and says Anthropic declined to change the protocol, describing the behaviour as expected.
Publishers:ox.security
Reality
- Evidence32
- Adoption34
- Hype gap+46
- Incentives86
- Confidence33
A Black Hat USA 2026 keynote described AI tooling that found roughly 1,000 bugs and then stalled on reporting. Patch Tuesday volume tells the same story from the other end.
Reality
- Evidence30
- Adoption24
- Hype gap+34
- Incentives54
- Confidence30