One Article Review
| Source |  knowbe4 |
|---|---|
| Identifiant | 8332823 |
| Date de publication | 2023-05-02 13:00:00 (vue: 2023-05-02 13:07:04) |
| Titre | Cyberheistnews Vol 13 # 18 [Eye on Ai] Chatgpt a-t-il la cybersécurité indique-t-elle? CyberheistNews Vol 13 #18 [Eye on AI] Does ChatGPT Have Cybersecurity Tells? |
| Texte |
CyberheistNews Vol 13 #18 | May 2nd, 2023
[Eye on AI] Does ChatGPT Have Cybersecurity Tells?
Poker players and other human lie detectors look for "tells," that is, a sign by which someone might unwittingly or involuntarily reveal what they know, or what they intend to do. A cardplayer yawns when they\'re about to bluff, for example, or someone\'s pupils dilate when they\'ve successfully drawn a winning card.
It seems that artificial intelligence (AI) has its tells as well, at least for now, and some of them have become so obvious and so well known that they\'ve become internet memes. "ChatGPT and GPT-4 are already flooding the internet with AI-generated content in places famous for hastily written inauthentic content: Amazon user reviews and Twitter," Vice\'s Motherboard observes, and there are some ways of interacting with the AI that lead it into betraying itself for what it is.
"When you ask ChatGPT to do something it\'s not supposed to do, it returns several common phrases. When I asked ChatGPT to tell me a dark joke, it apologized: \'As an AI language model, I cannot generate inappropriate or offensive content,\' it said. Those two phrases, \'as an AI language model\' and \'I cannot generate inappropriate content,\' recur so frequently in ChatGPT generated content that they\'ve become memes."
That happy state of easy detection, however, is unlikely to endure. As Motherboard points out, these tells are a feature of "lazily executed" AI. With a little more care and attention, they\'ll grow more persuasive.
One risk of the AI language models is that they can be adapted to perform social engineering at scale. In the near term, new-school security awareness training can help alert your people to the tells of automated scamming. And in the longer term, that training will adapt and keep pace with the threat as it evolves.
Blog post with links:https://blog.knowbe4.com/chatgpt-cybersecurity-tells
[Live Demo] Ridiculously Easy Security Awareness Training and Phishing
Old-school awareness training does not hack it anymore. Your email filters have an average 7-10% failure rate; you need a strong human firewall as your last line of defense.
Join us TOMORROW, Wednesday, May 3, @ 2:00 PM (ET), for a live demonstration of how KnowBe4 |
| Envoyé | Oui |
| Condensat | #18 #18 000+ 1767 1848 1921 1994 2008 2018 2022 2023 2023:https://blog 21/ 2:00 2nd 30+ 356bd1a3 576 6th 7th :https://explore :https://users :https://www :https://youtu about abus accept accomplish according accordingly account accounts actions active actors actually adams adapt adapted add added adding additional adjusts ads advanced advances africa afterwards against ago ahead ai:https://www alert alexei all already also always amazon amount angela announced answers any anymore anything anytime anywhere apologized: app apparently appear appears approach appropriate april are argues arizona arthur article artificial asimov ask asked asks assist assume atlantic: attachment attachments attack attacks attacks:https://www attend attention audio author authorities authors automated auxiliaries auxiliary available average avertium award awareness baby bank based bbc be/4hkxofq6jme be/8qwaa46utkm be/mq3dkhmjwxc be/spcqrd be/vkgys71n72c be/wfzshrpqr38 be/wlw3vvua be/wubs3xyfroo be/zwhl9dh because become been before being believed believing benchmarking best betraying bigger biggest bike binance blog bloghttps://blog bloomberg bluff boris both box brands breach breaches bridge brother bruce buckle bull bungee business businesses busy but buttes button call can candidate cannot car card cardplayer care case cases cat:https://youtu catching caution center centers: central ceoknowbe4 chair chairman challenging chancellor change; changes changing chatbot chatgpt check chief china chn christine civilian civilian/ claims class classified clear click cloned cmu com/article/3694850/iranian com/article/3694931/5 com/articles/i com/chatgpt com/cyberheistnews com/email com/eyes com/flying com/knowbe4 com/missing com/news/cyberattacks com/one com/phishing com/qbot com/resource/can com/stories/5928f4b47ca643929516331e6638596b/ukraine com/story/nsa com/technology/2023/04/nsa com/technology/archive/2023/04/ai com/tv com/watch com/wcc/r/4194769/5203727bef3186811d42096b28caa603 come comes coming commenting common company compare competitors complains complete complimentary comprehensive compromise compromised computer concise conduct confidential confirm confront consider constantly consultant content content: context continue continued continues contracts contributions controls conversation conversations corporation correct could countries course coyote crawling create credentials crimes:https://www critical csoonline culture cumulative cunning current customers cyber cyberattacks cybercriminals cyberheist cyberheistnews cybersecurity cyberspies cyberwire d427 dangerous dark data databases date date/time: day ddos deal decades decoded dedicated deepfake defend defense defenseone deliver demand demo demonstration denial deploy depth destroying details detect detection detectors determines develop development devious did difficult dilate directed direction director directory discovery disinformation distributed does dolomite don donations down downloads doxing dozens drawn dream driven duo during earth easily east easy ece edu/~gamvrosi/thelastq education eec effectively efforts either election eliminating email emails emergence emoji emphasis end endpoints endure enforcement engaged engineer engineered engineering enhance enormous ensure environment equally erection ethical european evangelist even ever every evolves example exchange exchanged executed executive executives expectations expecting expects experience explained exposed exposure eye eyes face faceted fact failed failure fair falls familiar family famous far fashion fast fave favorite feature features fed federal fell fending field figure file files filters finally finance find firewall first fleets:https://www flixxy flooding flooding:https://youtu flying foe fooled footage forays foreign form former formidable fraud frequency frequent frequently fresh friendly from fun further future gangs gap generate generated generative genuine germany get getting girder goat going golden good gorgeous got gotchas government governmental governments gpt grain greg grimes group groups grow g |
| Tags | Ransomware Malware Hack Threat |
| Stories | ChatGPT ChatGPT |
| Notes | ★★ |
| Move | 
|
L'article ne semble pas avoir été repris aprés sa publication.
L'article ne semble pas avoir été repris sur un précédent.