ai人工智能的本质和未来_人工智能手中的网络安全的未来

ai人工智能的本质和未来

A recent Synack Report claims that combining cybersecurity talent and AI-enabled technology results in 20x more effective attack surface coverage than traditional methods.

Synack的最新报告声称，将网络安全人才和支持AI的技术相结合，可以使攻击面的有效覆盖范围是传统方法的20倍。

But it’s difficult to truly understand the implication of these numbers. Most content on the topic leaves the reader to do all the math, connect the dots, and try to understand the real problem behind the numbers, all by themselves — an overwhelming task.

但是，很难真正理解这些数字的含义。 该主题的大部分内容都使读者自己去做所有的数学运算，连接点，并试图独自理解数字背后的实际问题，这是一项艰巨的任务。

Two of the most interesting topics on the internet today are Artificial Intelligence and Cybersecurity. You want solid takeaways and insights to embrace in today’s inter-connected digital economy.

当今互联网上最有趣的两个主题是人工智能和网络安全。 您想要扎实的收获和见解，融入当今互联的数字经济中。

人工智能的挑战：网络安全行业的观点 (Challenges of AI: A cybersecurity industry perspective)

A recent survey by the Consumer Technology Association identifies AI’s top application as cybersecurity, with 44% of all AI applications being used to detect and deter security intrusions.

消费者技术协会(Consumer Technology Association)最近进行的一项调查将AI的顶级应用程序确定为网络安全， 所有AI应用程序中有44％用于检测和阻止安全入侵。

At the other end of the spectrum, cybercriminals are increasingly (and innovatively) using AI as digital ammunition. It is to create potent security threats, resulting in an increased demand for cybersecurity experts.

另一方面，网络犯罪分子越来越多地(并且以创新方式)将AI用作数字弹药。 它会产生强大的安全威胁，导致对网络安全专家的需求增加。

On the subject of threats, here’s my pick of the top-3 current concerns:

关于威胁，这是我目前最关注的三大问题：

In 2019, the Wall Street Journal reported that a group of cybercriminals had demanded €220,000 from the CEO of a U.K.-based energy firm. An interesting twist to the plot: they used AI to impersonate the CEO’s voice. The lesson learned? AI provides cybercriminals with an additional edge to bypass security blind spots in an organization.

2019年，《华尔街日报》 报道称 ，一群网络犯罪分子向一家英国能源公司的首席执行官索要22万欧元。 情节的一个有趣转折：他们使用AI模仿了CEO的声音。 吸取的教训是？ 人工智能为网络犯罪分子提供了额外的优势，可以绕过组织中的安全盲点。

2.网络安全攻击和数据泄露的数量和复杂性不断提高 (2. Accelerated volume and complexity of cybersecurity attacks and data breaches)

A study by Varonis puts things into perspective:

瓦罗尼斯(Varonis)进行的一项研究将其视为一种观点：

Long story short, AI-powered technology can enable automation of tasks, protect the attacker’s identity, spawn the next frontier in enterprise fraud, a.k.a deepfakes, and refine malicious services (to name but a few instances) at a rate and scale that’s virtually unheard of.

长话短说，基于AI的技术可以使任务自动化，保护攻击者的身份，在企业欺诈，akafakefake中产生下一个前沿领域，并以几乎闻所未闻的速度和规模完善恶意服务(仅举几例)的。

3.由于更多地依赖于人工智能技术，(威胁)机会之窗 (3. A window of (threat) opportunity, owing to greater reliance on AI-enabled technology)

Practically speaking, AI-led technology can swing both ways. Either in favor of organizations or in favor of cybercriminals, depending on the circumstances. Organizations can beef up their security by using emerging technologies and AI-led software.

实际上，以AI为主导的技术可以双向摇摆。 视情况而定，有利于组织还是有利于网络犯罪分子。 组织可以通过使用新兴技术和AI主导的软件来增强其安全性。

Ironically, attackers can overcome these security controls and manipulate routine tasks by using the same technology in an improvised capacity, giving rise to more complex and interconnected risks. Believe it or not, in 2018, Cisco reported that they “blocked seven trillion threats, or 20 billion threats a day, on behalf of their customers.” Currently, it seems like we’re stuck in a “chicken-egg” situation.

具有讽刺意味的是，攻击者可以通过临时使用相同的技术来克服这些安全控制措施并处理日常任务，从而带来更加复杂和相互关联的风险。 信不信由你， 思科在2018年报告说，他们“代表客户阻止了7万亿次威胁，或每天阻止200亿次威胁。” 当前，似乎我们陷入了“鸡蛋”的境地。

人工智能如何为网络安全流程做出贡献 (How AI is contributing to cybersecurity processes)

According to Capgemini, two out of three organizations are planning to adopt AI solutions by 2020.

凯捷(Capgemini)表示，三分之二的组织计划在2020年之前采用AI解决方案。

Despite posing serious threats, AI augments and complements the future of cybersecurity best-practices. Here’s a lowdown of the “AI advantage” in the cybersecurity landscape:

尽管构成了严重威胁，但AI增强并补充了网络安全最佳实践的未来。 以下是网络安全领域中“ AI优势”的缩影：

Synack claims that “using AI accelerates the time to evaluate the breach-worthiness of vulnerability by 73%.”

Synack声称“使用AI将评估漏洞的可破解性的时间缩短了73％。”

Plus, additional data by Capgemini research shows that “3 out of 4 executives believe AI in cybersecurity speeds up breach response — both in detection and remediation.

此外，Capgemini研究的其他数据显示，“四分之三的高管认为，网络安全中的AI可以加快对漏洞的响应-在检测和补救方面。

And around 64% said that it also reduces the cost of detection and response.” Expert Kayne McGladrey’s take on AI explains what these numbers might mean. Talking to CIO, she says:

大约有64％的人表示，这也降低了检测和响应的成本。” 专家Kayne McGladrey对AI 的看法解释了这些数字可能意味着什么。 与CIO交谈时，她说：

“AI’s tenacity results in reducing time to discovery, which doesn’t need holidays, coffee breaks, or sleep. And, is unlike Tier 1 security operations center analysts for whom reading endless log files and alerts gets boring.”

“ AI的毅力可以减少发现时间，不需要假期，喝咖啡或睡觉。 而且，与第1层安全运营中心的分析师所不同的是，他们无休止地读取日志文件和警报。

AI is like a superpower organization that can use in a number of ways — to review user behaviors, find patterns, and locate irregularities in the security network to name just a few.

AI就像一个超级大国组织，可以通过多种方式使用-审查用户行为，查找模式以及查找安全网络中的违规行为，仅举几例。

Image Source

图片来源

The use of AI in combination with human efforts can help companies find and close critical vulnerabilities 40% faster.

将AI与人工相结合可以帮助公司更快地发现和关闭关键漏洞， 速度提高40％ 。

Plus, this allows companies to focus their efforts on creating secure ‘choke points’ instead of spending millions to secure the entire work environment.

此外，这使公司可以将精力集中在创建安全的“阻塞点”上，而不是花费数百万美元来保护整个工作环境。

Finally, by using AI for improved security, companies can leverage greater returns and minimize risks — a growing-concern among companies today.

最后，通过使用AI来提高安全性，公司可以利用更大的回报并最大程度地降低风险-当今公司越来越关注这一点。

AI is often described as intelligent, and rightly so. There are numerous application areas in which this powerful technology is helping organizations highlight recurring incidents and reverse the damage. (More on this in the next section).

人工智能通常被描述为智能，这是正确的。 这项功能强大的技术可在众多应用领域中帮助组织突显重复发生的事件并消除损害。 (有关更多信息，请参见下一节)。

网络安全中AI的前三大影响用例 (Top-3 impactful use cases of AI in cybersecurity)

1.梳理大量安全数据并自动执行日常任务 (1. Combing through mountains of security data and automating routine tasks)

One of the biggest concerns in cybersecurity is the sheer volume of data organizations have to tackle daily. This is probably one of the reasons why Gmail chose the AI route to block 100 million extra spam messages every day CTO of Seedcamp, David Mytton’s explanation of how AI is disrupting the cybersecurity space is on-point.

网络安全中最大的担忧之一是组织每天必须处理的大量数据。 这可能是Gmail选择AI路由来每天阻止Seedcamp的CTO 额外发送1亿条垃圾邮件的原因之一， David Mytton对AI如何破坏网络安全领域的解释已刻不容缓。

In CIO Mytton says, “As more and more systems become instrumented, the problem shifts from knowing that ‘something’ has happened. This is to highlight that “something unusual” has happened.”

迈特顿( CIO Mytton)表示： “随着越来越多的系统投入使用，问题从知道发生了某些事情转移了。 这是要强调“发生了不寻常的事情”。”

Instrumented system refers to as — who has logged in and when? What was downloaded and when? What was accessed and when?

仪器化系统指的是-谁登录以及何时登录？ 下载了什么以及何时下载？ 访问什么以及何时访问？

2.减少错误的安全警报并捕获异常事件 (2. Reducing false security alarms and capturing unusual incidents)

AI intelligence, in conjunction with threat intelligence, can detect new security issues and resolve them, offering a threat detection rate of 95% as opposed to traditional antivirus software, “where the detection rate is only about 90%, meaning 10% of malicious samples are being missed.”

人工智能情报与威胁情报相结合，可以检测并解决新的安全问题，与传统的防病毒软件相比，其威胁检测率为95％，而传统的防病毒软件的威胁检测率仅为90％，意味着恶意样本的10％被错过了。”

In the same vein, Michael Overly, partner at Foley & Lardner LLP, talked to Digital Munition about the benefits of addressing these missed threats in a time-sensitive manner, saying:

同样，Foley＆Lardner LLP的合伙人Michael Overly向Digital Munition谈到了以对时间敏感的方式解决这些遗漏威胁的好处，他说：

“The hope is that these systems will minimize false alarms and insubstantial issues, leaving a much smaller set of ‘real’ threats to review and address.”

“希望这些系统能够最大程度地减少虚假警报和无关紧要的问题，只剩下很少的'真实'威胁可供审查和解决。”

3.通过提供预测功能和提高效率来增强万无一失的安全性 (3. Empowering foolproof security by offering predictive functions and improving efficiency)

One of the most innovative and effective applications of AI in thwarting security breaches is the use of biometric authentication. Tech giant, Apple, uses this method — commonly known as “Face ID.”

在阻止安全漏洞方面，AI最具创新性和最有效的应用之一是使用生物特征认证。 科技巨头苹果公司使用这种方法，通常称为“面部识别”。

The powerful face ID technology combines built-in infra-red sensors and neural engines to recognize the user. If Apple’s claims are to be believed, the benefits are vast, with “only a one-in-a-million chance of fooling the AI to open a device with another face.”

强大的人脸ID技术结合了内置的红外传感器和神经引擎来识别用户。 如果可以相信苹果的说法，那么它的好处是巨大的，“只有百万分之一的机会愚弄AI打开另一张面Kong的设备。”

Apart from providing additional layers of security, AI is making teams more efficient, according to 70% of security professionals, and it is eliminating as much as 55% of employees’ manual tasks.

据70％的安全专业人员称，除了提供附加的安全层外，AI还使团队更加高效，并且消除了多达55％的员工手动任务。

With the additional layer of security and helping the team pivot their energies towards solving more important tasks, AI also powers productivity. Helping with productivity reduces overall stress levels for everyone involved.

通过附加的安全层并帮助团队将精力集中在解决更重要的任务上，人工智能还可以提高生产力。 帮助提高生产率可以减轻所有参与人员的压力。

人工智能安全的主要观点和未来趋势 (Key perspectives and future trends in AI security)

We’ve spoken plenty about how AI works round-the-clock to keep an enterprise secure, but what about the safety and training of the technology itself? Experts and enterprises need to refocus their strategies on training AI — ML models to max the value of these systems. ML systems can be trained to learn from historical data and detect anomalies. They can allow companies to mitigate and manage cyberattacks efficiently.

关于AI如何全天候工作以确保企业安全的问题，我们已经谈了很多，但是技术本身的安全性和培训又如何呢？ 专家和企业需要将其策略重新集中在培训AI-ML模型上，以最大程度地发挥这些系统的价值。 可以训练ML系统从历史数据中学习并检测异常。 它们可以使公司有效地缓解和管理网络攻击。

From boosting the security defense to hyper-automating parts of the cybersecurity processes, AI will play an increasingly important role in preventing cyberattacks.

从提高安全防御能力到网络安全流程的超自动化部分，人工智能将在防止网络攻击中发挥越来越重要的作用。

Capgemini recently reported that AI “adoption is set to skyrocket, with almost two out of three (63%) of organizations planning to employ AI by 2020.”

凯捷(Capgemini)最近报告说，“人工智能的采用将激增，三分之二(63％)的组织计划到2020年采用人工智能”。

As mentioned earlier, AI is a technology that can be used to both defend and attack an organization’s digital defense systems.

如前所述，AI是一种可用于防御和攻击组织的数字防御系统的技术。

The need of the hour is for cybersecurity experts to proactively identify attacks (think: Spam email attempts, disabling critical infrastructure, among others) and defend against them.

网络安全专家需要一个小时来主动识别攻击(例如：垃圾电子邮件尝试，禁用关键基础设施等)并防御攻击。

人工智能的局限性和可能的​​解决方案 (Limitations of AI & possible solutions)

“ AI won’t solve all your security problems. Think of it as a way to advance the security posture, not as a silver bullet.” — Raja Patel, vice president, security products at Akamai Technologies.

“ 人工智能无法解决您所有的安全问题 。 认为这是提高安全状况的一种方法，而不是灵丹妙药。” — Akamai Technologies安全产品副总裁Raja Patel。

Image Source

图片来源

According to research by White Hat Security, “ 60% of security professionals are still more confident in cyber threat findings verified by humans over those generated by AI.”

根据White Hat Security的研究，“ 60％的安全专业人员对人类验证的网络威胁发现比对AI产生的发现更有信心。”

The top-4 areas where human intelligence trumps AI in the operational security process are the use of intuition, creativity, human experience, and frame-of-reference — evolved capabilities that AI is yet to demonstrate, let alone master.

其中，人类的智慧胜过AI在运行的安全性过程中的顶4个地区均采用直觉，创造力，人的经验和框架的引用的-进化功能，AI是仍然需要证明，更不用说掌握。

Evidently, the current cybersecurity climate at least has tipped the balance towards human capabilities. The solution? Augmenting human talent with AI’s technological prowess can be a reliable way forward. Aarti Borkar, Vice President at IBM Security, writes in Fast Company of a 360-degree solution enterprises can embrace:

显然，当前的网络安全环境至少已将平衡权转移到了人的能力上。 解决方案？ 利用AI的技术实力来增强人才可能是可靠的前进之路。 IBM Security副总裁Aarti Borkar在Fast Company中写道，企业可以接受的360度解决方案：

“One way to help prevent bias within AI is to establish cognitive diversity. The diversity in the computer scientists developing the AI model, the data feeding it, and the security teams influencing it.”

“防止人工智能内部偏见的一种方法是建立认知多样性。 开发AI模型的计算机科学家的多样性，为它提供数据的数据以及影响它的安全团队。”

Some of the other growing concerns with the use of AI have been well captured by Osterman Research:

Osterman Research已很好地抓住了其他一些有关使用AI的担忧：

Moving forward, let’s look at how enterprises can solve these limitations from a holistic perspective:

展望未来，让我们从整体角度看企业如何解决这些限制：

• The first step is for companies to invest in an experienced cybersecurity firm with seasoned professionals.
  第一步是让公司投资于拥有经验丰富的专业人员的经验丰富的网络安全公司。
• Test systems and audit your hardware as well as software to find and proactively fix security gaps.
  测试系统并审核您的硬件和软件，以发现并主动修复安全漏洞。
• Install — and constantly update — firewalls and other malware scanners that keep your systems secure.
  安装并不断更新防火墙和其他恶意软件扫描程序，以确保系统安全。
• Review the latest cyber threats and security protocols to prioritize risks and develop effective strategies.
  查看最新的网络威胁和安全协议以区分风险优先级并制定有效的策略。

人工智能驱动的未来网络安全：专家演讲 (AI-Powered cybersecurity in the future: Expert speak)

We’ve chalked out the numbers and done our analysis. But what do the experts think about the prevalence and relevance of AI in cybersecurity? Let’s look at the picture painted for us by some of the expert opinions collected in Forbes:

我们已经画出了数字并进行了分析。 但是专家如何看待AI在网络安全中的普遍性和相关性？ 让我们来看看《福布斯》中收集的一些专家意见为我们绘制的图片：

Sami Laine, Director of Technology Strategy at Okta, says: “We will see threat actors use deepfakes as a tactic for corporate cyberattacks, similar to how phishing attacks operate.

Okta技术战略总监Sami Laine表示：“我们将看到威胁行为者使用深造品作为 企业网络攻击的 一种策略 ，类似于网络钓鱼攻击的运作方式。

Phishing attacks are where the money is for cybercrooks, and they can wreak serious havoc on unsuspecting employees.

网络钓鱼攻击是网络骗子的金钱之源，它们可能对毫无戒心的员工造成严重破坏。

What this means is that organizations will need to keep validation technology up-to-date; the tools — to create deepfakes and to detect them- will be the same. So, it’ll be an arms race for who can use the technology first.”

这意味着组织将需要使验证技术保持最新状态 ； 创建深造和检测深造的工具将是相同的。 因此，对于谁先使用这项技术将是一场军备竞赛。”

Audra Simons, Director of Innovation at Forcepoint, offers an interesting perspective: “We expect deepfakes to make a notable impact across all aspects of our lives in 2020 as their realism and potential increases.

Forcepoint创新总监Audra Simons提出了一个有趣的观点：“随着深度假货的真实性和潜力的增加，我们希望深度假货在2020年对我们生活的各个方面产生显着影响。

We will see Deepfakes-As-A-Service move to the fore in 2020 as deepfakes become widely adopted for both fun and malicious reasons.”

我们将看到Deepfakes-As-A-Service在2020年脱颖而出， 原因是出于娱乐和恶意原因 ，Deepfakes被广泛采用。”

Marcus Fowler, Director of Strategic Threat at Darktrace explains what it takes to fight AI with AI: “The building blocks are well in place for the rise of AI-powered cyberattacks in 2020, as more sophisticated defenses and access to open-source AI tools incentivize adversaries to supercharge their attacks.

Darktrace的战略威胁总监Marcus Fowler解释了与AI对抗AI的方式：“随着2020年以AI为动力的网络攻击的兴起 ，更完善的防御措施和对开源AI工具的访问已到位激励对手加强攻击。

AI won’t only enable malware to move stealthily across businesses without requiring a human’s hands on the keyboard, but attackers will also use AI in other malicious ways, including determining their targets, conducting reconnaissance, and scaling their attacks.

人工智能不仅可以使恶意软件在企业中秘密移动，而无需人工操作，而且攻击者还将以其他恶意方式使用人工智能，包括确定目标，进行侦察和扩大攻击范围。

Security experts recognize that defensive AI is the only force capable of combating offensive AI attacks. And that the battle must be fought by matching — or exceeding — the speed with which attackers innovate.”

安全专家认识到， 防御性AI是唯一能够抵抗攻击性AI攻击的力量。 而且必须通过使攻击者的创新速度达到甚至超过这一速度来进行战斗。”

Pascal Geenens, Security Researcher at Radware talks about this hot topic plaguing organizations and countries across the globe: “ Disinformation and fake news can spread havoc in both the public and private sector and is increasingly being used as a weapon by nation-states.

Radware的安全研究员Pascal Geenens谈到了这个困扰全球组织和国家的热门话题：“ 虚假信息和虚假新闻可能在公共和私营部门中肆虐，并且越来越多地被民族国家用作武器。

In 2020, deep learning algorithms can bring about in generating fake, but seemingly realistic images and videos. This application of AI will be a catalyst for large scale disinformation campaigns. “

2020年，深度学习算法可以带来伪造但看似逼真的图像和视频的生成。 AI的这种应用将成为大规模信息传播活动的催化剂。 “

Phil Dunkelberger, President, and CEO of Nok Nok Labs, sums up how increasing digitization will impact organizations, cybercriminals, governments, and the world at large: “As digitization continues in 2020, data will become more valuable than ever before. Information that may have previously seemed trivial to the everyday consumer will hold significant value for stakeholders and hackers across the spectrum.

Nok Nok Labs总裁兼首席执行官Phil Dunkelberger总结了数字化的增长将如何影响组织，网络犯罪分子，政府和整个世界：“随着2020年数字化的持续发展，数据将比以往任何时候都更有价值。 以前对于日常消费者而言似乎微不足道的信息将对整个领域的利益相关者和黑客具有重大价值。

Adversaries or real-life ‘data bounty hunters’ will hunt for new ways to exploit it, governments will seek better ways to access it, enterprises will adopt stronger security measures to protect it, and end-users will demand better privacy to secure their personal information.

对手或现实生活中的“数据赏金猎人”将寻找新的利用方式，政府将寻求更好的方式来使用它，企业将采取更强的安全措施来保护它，最终用户将要求更好的隐私权来保护其个人安全信息。

Furthermore, with the rise of AI and machine learning, crucial data that impacts how medical decisions are made, where/how autonomous cars move. And more will become increasingly more mainstream-and increasingly more lucrative to threat actors pining for the information.”

此外，随着AI和机器学习的兴起，影响医疗决策方式，自动驾驶汽车在何处/如何移动的关键数据。 而且，越来越多的东西将成为威胁寻求信息的威胁参与者的越来越主流。”

网络安全的第一道和最后一道防线 (The first and last line of defense in cybersecurity)

Whichever way you slice it, one key learning, in particular, emerges from all this.

无论采用哪种方式切分，从这一切中都会特别获得一项关键的学习。

“AI isn’t ready to fly solo any time soon.”

“人工智能还没有准备好在不久的将来单飞。”

Enterprises, as well as experts, are more comfortable embracing a “middle-ground” approach. AI can be used as a smart tool to augment human intelligence. And help organizations stay competitive and safe in the ever-expanding world of cybersecurity threats.

企业和专家都更乐于采用“中间立场”方法。 人工智能可以用作增强人类智能的智能工具。 并帮助组织在网络安全威胁不断扩大的世界中保持竞争力和安全性。

In a nutshell, the advantages of AI far outweigh the limitations and offer a hopeful (and secure) way forward. What are your thoughts?

简而言之，人工智能的优势远远超过了局限性，并提供了一种有希望的(安全的)方法。 你觉得呢？你有没有什么想法？

Laduram Vishnoi

Laduram Vishnoi

The Future of Cybersecurity in the Hands of AI was originally published on ReadWrite on August 6, 2020 by Laduram Vishnoi.

人工智能手中的网络安全的未来 最初 由 Laduram Vishnoi 于2020年8月6日 发布在 ReadWrite 上 。

翻译自: https://medium.com/readwrite/the-future-of-cybersecurity-in-the-hands-of-ai-766b7fb6d5ea

ai人工智能的本质和未来