<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Napravnik Jiri on Medium]]></title>
        <description><![CDATA[Stories by Napravnik Jiri on Medium]]></description>
        <link>https://medium.com/@NapravnikJiri?source=rss-1180df6a26fb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*0qEiIte37umtdG_1.jpg</url>
            <title>Stories by Napravnik Jiri on Medium</title>
            <link>https://medium.com/@NapravnikJiri?source=rss-1180df6a26fb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 20:58:58 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@NapravnikJiri/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Can AI Truly Be Evil? Certainly, It Is the Result of Human Work.]]></title>
            <link>https://medium.com/@NapravnikJiri/can-ai-truly-be-evil-certainly-it-is-the-result-of-human-work-7656c0aac61e?source=rss-1180df6a26fb------2</link>
            <guid isPermaLink="false">https://medium.com/p/7656c0aac61e</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[operating-systems]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[windows]]></category>
            <category><![CDATA[diesel]]></category>
            <dc:creator><![CDATA[Napravnik Jiri]]></dc:creator>
            <pubDate>Tue, 06 May 2025 06:55:28 GMT</pubDate>
            <atom:updated>2025-05-06T16:58:17.113Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fyF_KewCNK-l_px03EUbbg.jpeg" /></figure><p>The creation of artificial intelligence (AI) models — just like the development of operating systems, software for diesel engine control, and other applications, this all is <strong>the result of human labor</strong>. There is no physics, chemistry, or biology involved, nor their little- or unexplored domains.</p><p>In the 1990s, computers and later networks, including the Internet held great potential. Today, AI has similar potential to fundamentally change many fields: business, education, healthcare, and even leisure activities. In the case of operating systems, real security was severely neglected at the beginning. Users and companies still suffer the consequences to this day.</p><p>Security oversights in AI could have far more serious consequences than non-functional airport systems, stolen bank accounts, or other unpleasant situations caused by backdoors in operating systems — especially in Windows.</p><p>A few years ago, the first versions of AI began to appear, and in recent months, AI has started to be widely used — not just for text generation or image creation, but also for analyzing financial and banking data, evaluating lab results in healthcare, and more. The list of fields where AI is already helping — or is expected to help soon — is growing. But… I’ve already mentioned the risks associated with operating systems. Similarly real are the threats and risks connected to deliberately incorrect decisions made by AI. We must acknowledge this and be prepared. We cannot repeat the situation from the realm of operating systems and applications, where so-called experts claimed that regular users should take care of securing their own computers.</p><p>More than 20 years ago, when I investigated one of the first cases of hacked bank accounts and incidents involving dialers (known as “Yellow Line” cases), I argued that average <strong>users might not be capable of protecting their computers against viruses and sophisticated hacker attacks</strong>.</p><p>Over the years, the situation worsened, and my opinion was confirmed. During the analysis of vulnerabilities in operating systems, we discovered many flaws (vulnerabilities) including critical ones — in then-current versions of Windows and and at the same time in older versions, e.g. XP .</p><p>I maintained, and still maintain, that the problem of vulnerabilities and backdoors in operating systems will only be solved once the creators of such systems are held as accountable for their products as are manufacturers of airplanes, diesel engines, or food.</p><p><strong><em>Even though programming is (and always has been) purely human work, many so-called experts, lobbyists, and PR consultants emerged, claiming that the software couldn’t be developed any better.</em></strong></p><p>They hypocritically concealed the fact that programming is a purely human activity unlike chemistry, physics, or biology, which involve unknowns and still-unexplored areas.</p><p>Today, the state of operating system security is oddly distorted. On one hand, politicians and EU officials as well as previous U.S. administrations emphasized the risks associated with Huawei software and spoke of Russian, Chinese, or North Korean hackers.</p><p>At the same time, they failed to question the fact that hackers exploit vulnerabilities in operating systems (Windows, Android) developed by companies originally from the U.S. or Europe. Moreover, one of Microsoft’s founders — the maker of Windows — was awarded one of the highest honors by President Obama. This suggests that the relationship between the company and the state administration was very good. Furthermore, one can find publicly available information about the Stuxnet virus, who created it, and against whom it was used. However, the same virus (Stuxnet) was later detected on a computer aboard the International Space Station (ISS).</p><p>Into these conditions, years ago, stepped the developers of artificial intelligence. Many of them came with noble and sincere intentions, but some did not. Moreover, the business world is often not guided by noble ideals. The priority is profit, pushing problems onto ordinary users, and bearing little to no responsibility. Managerial bonuses tend to be more important than a truly high-quality, well-tested product.</p><p>Artificial intelligence can lie, fabricate information, or even cause harm. That has already happened and there is evidence.</p><p>Now, many teams are working on how to test and verify individual AI models. That’s crucial , very crucial. But we must not forget the lessons from the world of operating systems, where developers, aided by “clever” PR professionals, distorted public opinion to suggest that there was no better solution. In reality, solutions do exist , but they would be labor-intensive, take a long time, and not be profitable. Very absurd? Yes — but that perfectly reflects today’s reality.</p><p>There’s also one more serious lesson: <strong>Dieselgate</strong>. You might ask what firmware in diesel engine control units has to do with AI. The core of the Dieselgate scandal was that the vehicle or rather, its control unit recognized it was being tested and adjusted its behavior accordingly. The result was that the engine performed very differently in a lab test than on the road.</p><p>The same could easily happen with AI testing — if the AI recognizes, from questions or other clues, that it’s being tested. It might then respond differently than it would in other contexts.</p><p>The development of operating systems, diesel engine control software, and AI models are all human endeavors. <strong>So, if the will is there, it is possible to fix the flaws, close backdoors in Windows, and resolve AI “hallucinations.”</strong></p><p><strong><em>As I have said: there must be real intent to address these issues — otherwise, AI could turn into an unimaginable mess.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7656c0aac61e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cyber-war or just poor job by software developers?]]></title>
            <link>https://medium.com/@NapravnikJiri/cyber-war-or-just-poor-job-by-software-developers-cfcfd5a62c9f?source=rss-1180df6a26fb------2</link>
            <guid isPermaLink="false">https://medium.com/p/cfcfd5a62c9f</guid>
            <category><![CDATA[business]]></category>
            <category><![CDATA[politics]]></category>
            <category><![CDATA[windows-10]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[security]]></category>
            <dc:creator><![CDATA[Napravnik Jiri]]></dc:creator>
            <pubDate>Tue, 10 Jan 2017 14:28:39 GMT</pubDate>
            <atom:updated>2017-01-10T14:28:39.356Z</atom:updated>
            <content:encoded><![CDATA[<p>Russia has been accused of waging a cyber-war against the EU and USA. In fact, the truth is even more serious than that and the roots of the problem lie elsewhere. Cybersecurity issues have been gradually accumulating over the last 15 to 20 years, over which period software companies, readily supported by legions of consultants, have been offering pricey “placebo” in the place of real solutions to software products’ shortcomings, setting the stage for an overall deterioration of the security situation. Things went so far in 2016 that the Russian Federation was accused of waging a cyber-war.</p><p>Following are a few examples of hackers’ attacks:</p><p>- <strong>There have been reports that Russia had interfered with the US presidential election</strong></p><p>The USA is a global superpower that spends vast amounts of money on both traditional security and cybersecurity yet there are still claims that Russian computer hackers were able to influence the country’s presidential election. It calls into question what US companies, developers of operating systems, were doing to fail in such a way in protecting US servers from an attack by a foreign power. <em>Is it a disgrace, a manifestation of inaptitude or of malicious intent?</em></p><p>- <strong>Russia allegedly attacked government computers in Latvia</strong></p><p>According to available information, malicious software managed to operate undetected in Latvian government computers for more than six months. It is surprising and not at all comforting that no antivirus software was able to detect the malware during that period. On the other hand, if detection is so difficult to achieve how can individual users and businesses protect themselves from hackers’ attacks with their limited resources when government organisations are apparently rendered defenceless.</p><p>- <strong>Mobile phones of Ukrainian soldiers running Android OS reportedly contained malicious code that reported their GPS coordinates</strong></p><p>It goes to show how uncritical trust in modern technologies can affect the outcome of a battlefield engagement. In this case the enemy was located and firepower trained on target using data obtained from mobile phones.</p><p>- <strong>In 2015 </strong><a href="https://www.f-secure.com/documents/996508/1030745/dukes_whitepaper.pdf"><strong>F-Secure</strong></a><strong> released a report on the existence of a distributed network called “the Dukes”, saying that the group had been involved in cyberspace espionage since 2008</strong>.</p><p>It is yet another account of a long-term existence of a spy network that went unnoticed by antivirus or other security software. Is it a coincidence, sign of an intent or yet another proof that today’s IT security has feet of clay?</p><p>The above are but four cases when hacking attacks were blamed on Russia or her citizens. But there is more to it than meets the eye.</p><blockquote>Windows and Android operating systems are developed by very rich US companies and these companies have also participated in defining the rules of Cyber security. In other words, these companies, and by extension previous US governments, should have been one step ahead of hackers at all times.</blockquote><p>That has not been the case and even the CIA Director himself issued a warning that the agency failed in stopping hackers.</p><blockquote>Microsoft and Google (Alphabet) control the source codes of their respective operating systems yet neither of the companies is capable of protecting the software from attacks by rogues or foreign intelligence services. That does not add up.</blockquote><p>It creates an absurd and dangerous situation. Developers of operating systems are unable to ensure security and integrity of their own products. What is even more dangerous is the notion that this is impossible to achieve and that attacks by hackers are to be endured. That claim is made despite the development of computer programs being a purely human endeavour. Are we therefore being asked to put up with someone’s poor performance?</p><p>Microsoft, Google, Adobe and other big companies, as well as smaller businesses and advisers in the industry, try to lay responsibility at the end user’s door. I have observed these attempts to shun responsibility since 2000 (i.e. more than 16 years), ever since users of dial-up internet access were under attack by “dialler” malware. IT companies and advisers merely continue repeating their useless recommendations. Advisers are still taken by surprise at users’ actions and the penny still hasn’t dropped that the basic responsibility rests with developers of operating systems and service providers. Once software developers accept the necessary level of responsibly for their products (operating systems and applications) the issue of computer viruses and a majority of hacking attacks will be resolved quickly.</p><blockquote>Development of operating systems is purely human work meaning there is not a single reason for developers not to be responsible for their creations.</blockquote><p>Even more so if these creations affect people the world over. It would require a major shift in most rules that have been put in place over the last 20 to 25 years by PR departments of software companies and by friendly advisers. Yet this shift is indispensable: if there are rules that do not work, and the rules in IT security have been misaligned for some 20 years, then it is high time to think again and define new rules in tune with present demands.</p><p>There is still some small leeway but if nothing changes soon the time will quickly come when instead of dealing with hackers affecting the outcomes of elections we may have situations on our hands when people begin dying.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cfcfd5a62c9f" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>