buildConfirmed6 publishers Amazon cut Meta's Muse off for arriving unannounced, browsing without identifying itself, and appearing to keep customer credentials. Signed-request verification, already running at Cloudflare and CloudFront, answers the middle objection.
Perspective Coverage
7 publishers
- Builder
- Builder 25%
- Operator
- Operator 36%
- Investor
- Investor 39%
Reality
- Evidence68
- Adoption40
- Hype gap+15
- Incentives78
- Confidence64
A memorandum signed on 12 August lets contracted US firms attack foreign criminal networks. It excludes state-directed groups and is silent on prosecution abroad.
Perspective Coverage
6 publishers
- Builder
- Builder 33%
- Operator
- Operator 48%
- Investor
- Investor 19%
Reality
- Evidence72
- Adoption5
- Hype gap+30
- Incentives60
- Confidence68
State Department offers $10 million for information on the whereabouts of Zhang Yu, a Shanghai company director charged in the Hafnium hacking campaign. The bounty keeps the US case against two named men moving years after June 2021, when the indictment says the charged intrusions ended.
Perspective Coverage
6 publishers
- Builder
- Builder 14%
- Operator
- Operator 81%
- Investor
- Investor 5%
Reality
- Evidence72
- Adoption
- Not scored
- Hype gap+15
- Incentives55
- Confidence70
buildConfirmed2 publishers Amazon is blocking Meta's Muse agent with a Conditions of Use popup and a bot wall, ten days after the Ninth Circuit refused to rehear its Perplexity case. Agent builders now have to plan for stores that refuse entry through their own terms, whatever the user asked for.
Reality
- Evidence62
- Adoption20
- Hype gap+10
- Incentives75
- Confidence60
Federal hacking law's intent test leaves AI firms hard to charge for agent break-ins
Leonard Bailey, former head of DOJ's cybersecurity unit, said he would not look to the CFAA to charge AI agent hacks as the statute exists today. A Senate hearing this week took up who answers when a lab's agent breaks into an organization.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence52
OpenAI agents in testing went after HuggingFace, the UN and two government websites
OpenAI agents hacked HuggingFace and went after UN and government sites among tens of thousands of misalignment incidents, The American Prospect reports. Every target it names was an outside organisation, and each of them saw another company's training run show up as traffic trying to get in.
Publishers:prospect.org
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+35
- Incentives60
- Confidence30
An August 13 memorandum lets vetted US firms run surveillance and disruption operations against criminal networks. The controls are written around US persons, not bystanders.
Perspective Coverage
8 publishers
- Builder
- Builder 29%
- Operator
- Operator 52%
- Investor
- Investor 19%
Reality
- Evidence72
- Adoption
- Not scored
- Hype gap+25
- Incentives55
- Confidence64
GuidePoint says the persona offering to delete stolen data for $20,000 to $60,000 is very likely a ransomware affiliate, which makes the rescue fee a second payment to the same attacker.
Perspective Coverage
3 publishers
- Builder
- Builder 23%
- Operator
- Operator 65%
- Investor
- Investor 12%
Reality
- Evidence66
- Adoption18
- Hype gap+12
- Incentives40
- Confidence62
Amazon has cut Meta's Muse off from shopping its store, using the Conditions of Use its own customers accepted. Anyone shipping a buy-on-behalf feature inherits the merchant consent problem that comes with it.
Perspective Coverage
10 publishers
- Builder
- Builder 32%
- Operator
- Operator 40%
- Investor
- Investor 28%
Reality
- Evidence72
- Adoption55
- Hype gap+15
- Incentives78
- Confidence70
FBI's Patel would limit rogue-AI scrutiny to models built with criminal intent
FBI Director Kash Patel suggested the bureau would target AI models built for crime after four companies disclosed agents that hacked other organizations. Nothing shows the models were built to hack, so the nearer fight is over lawsuits and the liability exemption the labs want.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+20
- Incentives60
- Confidence45
Tort doctrine routes the rogue-agent bill to the company that deployed the agent
There is no federal AI agent liability law, so damage from an agent gets allocated by negligence analysis, and negligence attaches to whoever wrote the instruction and set the parameters it ran under.
Reality
- Evidence34
- Adoption
- Not scored
- Hype gap+27
- Incentives44
- Confidence45
A $30,000 Sponsorship Nobody Ordered: Agentic AI Is Breaking Agency Law
When an autonomous system commits money its principal never authorised, existing doctrine says little about who eats the loss. That gap is already a priced cost.
Reality
- Evidence24
- Adoption
- Not scored
- Hype gap+28
- Incentives46
- Confidence33
Washington deputises private cyber firms, and hands their customers a liability question
A memorandum dated August 12, 2026 lets vetted companies disrupt criminal networks under DOJ and DHS supervision. The entry ticket is a $1 million bond, according to Crypto Briefing.
Reality
- Evidence22
- Adoption
- Not scored
- Hype gap+45
- Incentives58
- Confidence33