Join daily and weekly newsletters to obtain the latest updates and exclusive content to cover the leading artificial intelligence in the industry. Learn more
They are called the new tradecraft opponents to exploit any weakness that can find them at the end points, relying on AI (GEN AI) to create new preferred attack weapons.
What causes anxiety is how quickly their graphics grow. This is evident in the speed and size of hunting campaigns, DeepFake videos, and social engineering attacks. more 67 % Hunting attacks depend on artificial intelligence last year, and 61 % One of the security leaders see the widely created hunting campaigns with AI Chatbots attacking their organizations. Deluette Fake’s deep losses are expected to rise $ 40 billion by 2027It grows at an annual growth rate of 32 %.
The cybersecurity teams that succeeded in adopting the attacks of the Venturebeat point that it is common to be the opponent that the survey would perform months before the attack to determine the weaknesses at the end points.
All it takes is a quick phone call to the internal service office to get a password or recycling MFA in a timely manner, and it is present.
The end points facing an attack from new attacks on artificial intelligence
The opponents offer priority for rapid attacks on end points using every source available for automation to expand their efforts, with GEN AI and machine learning (ML) preferred attack techniques.
Financial services, health care, manufacturing, distributors and basic companies in complex supply chains are the basic goals. Creating chaos in the supply chain of financial services is a double ransom.
“Because of the nature of our business, we face some of the most advanced and continuous electronic threats there,” Catherine Moin, Price companies’ SVP information security, tell Venturebeat in A. A recent interview. “We have seen others in the mortgage manufacturing, so we needed to make sure that this does not happen to us. I think what we are doing now is to fight artificial intelligence with artificial intelligence.”
The numerical -based weapons on the basis of Amnesty International have become so advanced that the violation could happen for several months without seeing a security team in the organization. The average time it takes to determine and contain a breach 277 days, with 176 days to get to know it and 82 days to contain it, based on The latest IBM cost for data breach report. AI weapons make it difficult for companies to bridge this gap.
“If you have two opponents erupted in two minutes, and it takes one day to receive data and another day to run the search, how can you hope to keep up with such an opponent?” Elia Zetsif, chief technology official in CrowdstrikeHe said Venturebeat newly.
One in three organizations does not have a documented strategy to defend the threats of artificial intelligence and GEN AI. Evante 2024 Cyber Security Report I found that 89 % of CISO and senior IT leaders believe that Amnesty International’s threats have just started.
The majority of security leaders, 60 %, Fear that their organizations are not ready to defend the threats and attacks operating in Amnesty International. IVANTI research has found that hunting, security gaps, ransom attacks, and API weaknesses are the four most common threats. It is not by chance that these four methods see the greatest gains from Gen Ai.
Protection of the end point needs more speed
“The opponent is increasing faster, and taking advantage of artificial intelligence technology is part of it. Take advantage of automation is also part of it, but entering this new wish is another important factor, and this is not only for modern attackers but also modern attack campaigns faster.”
ETay Maor, Senior Security Strategy in Cato networksI noticed during a recent interview to address that CATO networks already see cases “where attackers are trying to defraud the systems based on artificial intelligence by giving them quick injection, or not necessarily (S), but they are injecting information in the artificial intelligence system and trying to persuade that what is seen is not Harmful, it is benign. “
MAOR continued, “We participate and watch in various underground forums and see hundreds of artificial intelligence applications appeared. I think the institutions do not realize what is happening on their network, and the large headache will be once we see that harmful sliding through the cracks.”
Shailish Rao, President of Shalish Rao Palo Alto networks Dandruff section. “The attacks have become very sophisticated, and the needle changes billions of times a day. Do you prefer writing rules or applying machine learning on all these data?”
Vasu Jakkal, Vice President of Security, Compliance and Identity in Microsoft, I drew a more clear picture in breakEW With Venturebeat. “Three years ago in 2021, we saw 567 identity -related attacks, which were related to a password; these are many attacks per second. Today, this number is 7000 password attacks per second and more than 1500 followers followers.”
Four areas in which each provider of the end point needs to excel with artificial intelligence in 2025
The attacks of the end, identity and multiple fields dominate the threat system of the institutions today, which were partially fed by the new Tradecraft it invented using the Gen Ai.
The front point providers need to make progress in swallowing data, setting accident priorities, sorting and comforting, and analyzing the attack path. The finishing points suppliers include the pioneer who offer Cato Networks, CIOWDSTRIKE, Microsoft, Palo Alto Networks, Sentinelone, Trend Micro and ZSCaler, with Crowdstrike using AI and ML as basic ingredients Her trauma in 2011.
Here are four main areas, each seller needs to take action this year:
Speed to swallow data and normalization: The AI sellers help the scores for the records quickly from the end points, SAAS applications, local servers, and data drawing to a global scheme. This has the ability to reduce the time of analysis from days to minutes.
Improving identification of accidents and follow -up procedures: The correlation engines that work on artificial intelligence are keen on millions of alerts, and narrow them to a few high -value strands using the time chain data, IOA, and forms designated to determine the priorities of the most important accidents.
Speed up how the chandeliers of the end point platform and respond to infiltration attempts: AI’s tools help in advanced searches, create textual programs for treatment, and reduce manual forensic time from hours to minutes. The pre -created playing books allow quick procedures, such as isolating the end points or banning harmful IPS.
Enabling a more active position and improving the attack path analysis: Artificial intelligence determines potential infiltration methods by combining threat intelligence, weaknesses, user permits, and network data, then recommends targeted repairs to prevent multiple attack paths.
Play Book of 2025: 12 The gaps of artificial intelligence must be closed at the security of the end point
Amnesty International attacks with artificial intelligence should start to a more strategic level than currently in many organizations. It exceeds the extra end points with another agent, or users are required to ratify via multiple identity management systems. Artificial intelligence should be at the heart of cybersecurity.
The next twelve next is a pragmatic book of 2025, which covers the main technologies, processes and cultural transformations needed to fill the widen gaps in the security point of the end.
- Adoption of Sse or SSE: Adoption of the approximate Sse or SSE approach that mixes zero confidence with your network and the end and identity point. Let Amnesty International monitor everything in an actual time so as not to miss the threats that cannot see the tools.
- Modeling of a semantic data for a unified vision: Unifying records across the cloud, finishing points and identity systems in one model. Let artificial intelligence analyze data and normalize data until your team gets the full image quickly.
- Sorting and playing upper intelligence office: Use XDR or a similar system aligned with zero confidence to reduce housing times. AI’s play books help organize responses in minutes, not days.
- Motions -like engines to set threat priorities: Data are linked through zero confidence structure to capture surreptitious threats. Artificial intelligence can help suspicious patterns so that they can focus on real problems first.
- Prevention of the threat of identity: It depends on the principles of zero tenderness for the tests of the situation in actual time and concession analyzes. Artificial intelligence prevents attackers trying to determine stolen credentials or symbols.
- Clear stiffness by analyzing the attack path: Imposing zero confidence from the beginning to reduce the side movement. AI determines the lowest number of repairs that prevent multiple tracks in one corridor.
- Artificial intelligence and honor: It follows every decision that AI moves until the board of directors and the organizers trust. Zero Trust does not mean any black boxes. Maintaining vision in the logic of artificial intelligence.
- Use artificial intelligence specialized on general models: Training models on the real attacker’s tactics within the framework of zero work. You will see fewer misfortunes and the most accurate discovery.
- Continuous models synthesize and update the data collection: Update artificial intelligence models regularly to keep pace with advanced threats. Zero Trust is dynamic, so your data pipelines should also be.
- Checking human health in the episode: Even with the automation of zero confidence, human insight matters. Analysts improve the results of artificial intelligence to capture accurate threats and reduce wrong warnings.
- Automatically answering automatic accidents: Merging toys books from artificial intelligence with zero riches through end points, protection walls and identity. Once examined, the responses spread immediately.
- The integration of zero confidence from one side to the party: Check every step of the killing chain. Combining the discovery of artificial intelligence with strict access controls forcing the attackers to overcome new barriers at each turn.
The bottom line
Since the attackers go beyond the traditional end points, institutions must unify the threat data and accelerate their defenses through mixed infrastructure. This is why many of the main sellers need to accelerate their efforts by focusing on AI solutions that deal with swallowing data, relationship and automatic response in actual time.
The playing book above refers to reaching these goals and defense successfully against the hostile attacks based on artificial intelligence that will not only continue to come, but to continue to grow in development.
https://venturebeat.com/wp-content/uploads/2025/01/hero.jpg?w=1024?w=1200&strip=all
Source link