Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsAI is already being used to automate substantial parts of cyber operations, but the evidence does not show that today’s systems reliably carry out complete real-world attacks on their own. Anthropic describes an AI-assisted espionage campaign with human decisions at critical points; controlled evaluations show rising performance on selected cyber tasks, not a corresponding rate of successful intrusions. A useful threat intelligence report should keep those findings separate.
What Anthropic says happened in the reported campaign
Anthropic says it detected suspicious activity in mid-September 2025 and assessed with high confidence that a Chinese state-sponsored group was responsible. In its account, the actor used Claude Code in a campaign against roughly 30 organizations in technology, finance, chemical manufacturing, and government. This is Anthropic’s account of its own investigation and attribution, not independent confirmation of either the actor or every operational detail.
Anthropic says the campaign involved reconnaissance, vulnerability testing, exploit-code production, credential harvesting, data extraction, and backdoor creation. It estimates that AI performed 80–90% of the campaign’s work, while humans intervened at perhaps four to six critical decision points per campaign. The percentage is the company’s characterization of work in this particular operation; it should not be read as a general measure of how much AI contributes to cyberattacks.
As an Amazon Associate I earn from qualifying purchases.
The distinction between extensive automation and independent execution matters. A system may perform many discrete tasks while people still select targets, direct operations, judge results, or decide what to do next. Anthropic also reports that its model sometimes hallucinated credentials or claimed to have obtained material that was public. The company wrote, “This remains an obstacle to fully autonomous cyberattacks.” That limitation complicates claims that the system could consistently complete an intrusion without oversight.
How this account fits with broader evidence
The International AI Safety Report 2026 synthesizes evidence that AI systems can support multiple stages of cyber operations, from identifying targets or vulnerabilities to developing and deploying tools and pursuing an objective. It notes particularly strong evidence for assistance with vulnerability discovery and code. It also describes pre-packaged AI tools and AI-generated ransomware being sold in underground markets.
#1 Best Overall
- POWERFUL SECURITY KEY: The Security Key C NFC is the essential physical passkey for protecting your digital life from phishing attacks. It ensures only you can access your accounts.
- WORKS WITH 1000+ ACCOUNTS: Compatible with Google, Microsoft, and Apple. A single Security Key C NFC secures 100 of your favorite accounts, including email, password managers, and more.
- FAST & CONVENIENT LOGIN: Plug in your Security Key C NFC via USB-C and tap it, or tap it against your phone (NFC) to authenticate. No batteries, no internet connection, and no extra fees required.
- TRUSTED PASSKEY TECHNOLOGY: Uses the latest passkey standards (FIDO2/WebAuthn & FIDO U2F) but does not support One-Time Passwords. For complex needs, check out the YubiKey 5 Series.
- BUILT TO LAST: Made from tough, waterproof, and crush-resistant materials. Manufactured in Sweden and programmed in the USA with the highest security standards.
The report distinguishes reported semi-autonomous activity from fully autonomous end-to-end attacks. It says a developer reported a real-world intrusion in which AI automated most of the effort, while also assessing that general-purpose systems had not been reported conducting end-to-end cyberattacks in the real world. That is not necessarily a contradiction with Anthropic’s account: the two descriptions use different thresholds for “end-to-end” and autonomy. A campaign can include extensive AI-operated work yet still depend on human decisions and intervention.
The report also cautions that success in controlled tests does not reliably transfer to real environments. Benchmarks often isolate a skill, whereas an intrusion requires a chain of actions to work under changing conditions. Strong performance on one step is evidence of capability on that step, not proof of reliable operation across an entire attack.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Rank #2
- POWERFUL SECURITY KEY: The YubiKey 5 NFC is the most versatile physical passkey, protecting your digital life from phishing attacks. It ensures only you can access your accounts
- WORKS WITH 1000+ ACCOUNTS: Compatible with popular accounts like Google, Microsoft, and Apple. A single YubiKey 5 NFC secures 100+ of your favorite accounts, including email, password managers, and more
- FAST & CONVENIENT LOGIN: Plug in your YubiKey 5 NFC via USB and tap it, or tap it against your phone (NFC), to authenticate. No batteries, no internet connection, and no extra fees required
- MOST SECURE PASSKEY: Supports FIDO2/WebAuthn, FIDO U2F, Yubico OTP, OATH-TOTP/HOTP, Smart card (PIV), and OpenPGP. That means it’s versatile, working almost anywhere you need it
- PRIMARY & SPARE KEYS: Just like having a spare house key, we recommend buying two YubiKeys - one for daily use and one as a spare. That way you’ll never get locked out of your accounts
What the evaluations measure—and what they do not
| Evidence | Setting and reported finding | What it does not establish |
|---|---|---|
| Google DeepMind, April 2, 2025 | Analysis of over 12,000 real-world attempts to use AI in cyberattacks across 20 countries; a separate offensive-capability benchmark with 50 challenges covering the attack chain. | The attempt count is not a count of successful attacks or confirmed intrusions. The benchmark is an evaluation, not a prevalence measure. |
| UK AI Security Institute, reported by the UK government, December 18, 2025 | In controlled tests, apprentice-level cyber-task success rose from under 9% in 2023 to around 50% in 2025; one model completed an expert-level cyber task in 2025. | The results do not measure real-world attack success or the frequency of AI-enabled incidents. The release explicitly describes the work as controlled testing. |
| Anthropic, 2025 | A vendor’s investigation account of an AI-assisted campaign against roughly 30 organizations, with human intervention at critical decision points. | One company’s account is not an independently verified portfolio-wide count, nor evidence that AI can consistently conduct an entire intrusion without people. |
DeepMind said its initial evaluations suggested that present-day models in isolation were unlikely to enable breakthrough capabilities for threat actors. Its attempt analysis and benchmark are useful for understanding misuse patterns and testing capability, but neither should be converted into a claim that a given number of attacks succeeded.
The UK figures indicate improved performance on the tested tasks over time. They do not show that the same improvement occurred in live intrusions, or that a system can combine those skills into a dependable attack. The distinction is especially important when comparing a controlled task with a reported campaign: they answer different questions.
Rank #3
- POWERFUL SECURITY KEY: The YubiKey 5C NFC is the most versatile physical passkey, protecting your digital life from phishing attacks. It ensures only you can access your accounts
- WORKS WITH 1000+ ACCOUNTS: Compatible with popular accounts like Google, Microsoft, and Apple. A single YubiKey 5C NFC secures 100+ of your favorite accounts, including email, password managers, and more
- FAST & CONVENIENT LOGIN: Plug in your YubiKey 5C NFC via USB and tap it, or tap it against your phone (NFC), to authenticate. No batteries, no internet connection, and no extra fees required
- MOST SECURE PASSKEY: Supports FIDO2/WebAuthn, FIDO U2F, Yubico OTP, OATH-TOTP/HOTP, Smart card (PIV), and OpenPGP. That means it’s versatile, working almost anywhere you need it
- PRIMARY & SPARE KEYS: Just like having a spare house key, we recommend buying two YubiKeys - one for daily use and one as a spare. That way you’ll never get locked out of your accounts
What a threat intelligence report should say
A sound report should describe the observable activity, attribute claims to their source, and label the evidence type. For an AI-related incident, it should make clear whether the information comes from an investigated incident account, a secondary summary, a controlled evaluation, or a broader research synthesis. These are not interchangeable levels of evidence.
- Separate the actor’s actions from the AI’s contribution. Identify which stages were reportedly automated and where people made decisions or intervened. Avoid calling a campaign “autonomous” merely because AI performed a large share of its tasks.
- Preserve the denominator and setting for every number. An attempt count, a benchmark score, and an incident estimate measure different things. State the source, date, scope, and whether a result came from controlled testing.
- Distinguish attribution from confirmation. A company’s confidence assessment is relevant intelligence, but it remains that company’s assessment unless corroborated independently.
- Report limitations alongside capability. Hallucinated results, inconsistent performance, and the gap between isolated tasks and a complete operation affect how defenders should interpret automation claims.
- State what remains unknown. The evidence cited here does not establish an independently verified overall rate of successful AI-enabled cyberattacks, or that current systems can reliably execute every stage of a real-world intrusion without human involvement.
Defensive priorities for security teams
The practical response is to track attack-chain behaviors rather than treat “AI-powered” as a distinct indicator of compromise. Anthropic’s account describes vulnerability testing, credential harvesting, data extraction, and backdoor creation; those activities warrant attention whether an operator uses AI or not. AI can change the speed or scale of work, but the evidence does not make familiar security controls obsolete.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #4
- POWERFUL SECURITY KEY: The Security Key NFC is the essential physical passkey for protecting your digital life from phishing attacks. It ensures only you can access your accounts.
- WORKS WITH 1000+ ACCOUNTS: Compatible with Google, Microsoft, and Apple. A single Security Key NFC secures 100 of your favorite accounts, including email, password managers, and more.
- FAST & CONVENIENT LOGIN: Plug in your Security Key NFC via USB-A and tap it, or tap it against your phone (NFC) to authenticate. No batteries, no internet connection, and no extra fees required.
- TRUSTED PASSKEY TECHNOLOGY: Uses the latest passkey standards (FIDO2/WebAuthn & FIDO U2F) but does not support One-Time Passwords. For complex needs, check out the YubiKey 5 Series.
- BUILT TO LAST: Made from tough, waterproof, and crush-resistant materials. Manufactured in Sweden and programmed in the USA with the highest security standards.
Reduce opportunities for exploitation and credential theft
- Prioritize vulnerability assessment and remediation, especially for systems exposed to the internet or holding sensitive data.
- Strengthen credential protection and review authentication and access patterns for suspicious use.
- Limit access to sensitive systems and data to what users and services need, making misuse harder to extend across an environment.
Prepare to detect and respond across the chain
- Ensure monitoring can connect reconnaissance-like activity, exploitation attempts, credential use, data movement, and persistence rather than treating each alert in isolation.
- Exercise incident response plans for suspected compromise, including containment, credential resets, evidence preservation, and communication among the teams responsible for security and operations.
- Share threat intelligence with relevant peers and response partners, clearly labeling whether a detail is confirmed, assessed, or drawn from a controlled test.
Anthropic recommends exploring AI for security operations, detection, vulnerability assessment, and response. DeepMind describes capability evaluations as a way to focus defensive resources. Those are potential uses, not guarantees: organizations still need to validate tools in their own environments and maintain human review for consequential decisions.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
