Microsoft announced last week that it would be incorporating generative AI in everything, even security, as it had done with.NET many years ago.
In the.NET era, I joked about Microsoft being so out of control with.Net, that the bathrooms had been renamed Men.net or Women.net. Most of these efforts were not very logical. Microsoft’s decision to invest in generative AI now is more logical than when it was first implemented.
We’ll explore the impact of generative AI on security. We’ll end with the Product of the week: the BAC Mono, a custom-built street-legal track vehicle.
You are the biggest security risk.
We are often too excited by the technology available to us in order to minimize breaches. After layer upon layer of security software, it is still the case that people are the main cause of breaches. The majority of ransomware attacks and other problems are caused by someone tricked into revealing information.
Companies don’t practice them consistently or effectively. In that group I include executives from security companies who seem to believe the rules they helped to create do not apply to them.
When I was performing security audits on a CEO (who often boasted he was more knowledgeable about security than anyone in my division), I was able access his most confidential information, which was locked in a safe within 10 minutes. I did not use some James Bond-style hacking software, but instead looked in the secretary’s desk drawer that was unlocked.
Since decades, human error has been the leading cause of many of our worst security issues.
HP PC Security Solutions
This is being written at HP’s Amplify Partner Event, where HP has just launched its security solution. HP’s Wolf Security may be the best PC-security solution on the market.
HP stated that its security division generates 8 trillion dollars in revenue. This is only a fraction the amount of money it protects. All this technology is useless if it can’t stop an employee from making a stupid mistake.
HP’s technology offers VMs and BIOs as well as some of the best security solutions that I have seen. However, this only applies to someone who has accidentally lost or misplaced a computer. The HP technology doesn’t cover employees who intentionally or unintentionally breach their own security.
HP Sure Click is an exception, as it helps to prevent users from clicking on links they shouldn’t. Sure Click isolates potentially harmful actions within a virtual environment, so that they don’t cause damage to VMs. This is a great effort. HP is doing more than others, but it’s still not enough.
AI Security: Examples and Reasons for Needing It
I was once contacted by a CIO that was fired via email. He was so furious that he used the credentials he had been given to reformat every hard drive of his former company, effectively forcing them out. He was jailed and sued to poverty, but it didn’t help his former company.
In a second massive breach, the attacker used stolen credentials to gain uncontested access to an HR system of a company and sent a mass email to all non-management employees telling them that the firm was sold and they needed to give their banking details to receive their final check.
Almost every employee divulged their personal information before anyone thought to consult a manager. When the operation was stopped, the servers that were attacking were off-line, and the thieves had long since disappeared.
These exploits would have been successful even if HP’s Wolf Security had not been installed. The first was because there was no laptop involved, while the second was because it had been a phishing attempt that allowed access and compromise to an HR system.
HP is not the target here, because no other tech company or HP can effectively address a problem originating from employees. AI could potentially help with this “yet”.
BlackBerry and Microsoft: AI at the Rescue
Microsoft’s Security Copilot focuses initially on providing information to security professionals in real-time about current and potential breaches so that they can quickly mitigate them. This should address the problem of under-staffing and under-resourcing security. Most of these AI-generated initiatives are initially focused on increasing productivity and reducing employee burdens.
The real promise of generative AI lies in its ability to learn from employee behaviors and, by doing so, mitigate them. BlackBerry’s and Cylance units are the only companies that have aggressively fought against employee exposure using older AI technology.
BlackBerry’s technology monitors its employees, and it will block anyone behaving in an unusual way. For example, a service representative who suddenly begins downloading files related to the employee database or product development of a firm — indicating that someone was using that professional’s credentials.
Generative AI is capable of going much further, and moving more quickly. Generative AI, using huge models, can predict future behaviors, identify employees who violate company policies regularly (indicating that they are more likely act improperly), recommend remedial measures ranging from recurring automatic training to termination of employees most likely to cause a breach. This eliminates potential problems before an incident.
Before you become upset over the “termination”, realize that the remedy is not just termination, but may also include financial penalties or jail time depending on the severity and nature of the violation. Even if the employee is terminated, it’s still better than what would have happened otherwise.
Wrapping up: Generative AI, the Future of Security
Microsoft is the latest to bring AI into security. This could lead to the final elimination of one of our greatest security risks: humans. We will have the chance to finally mitigate the security issue that has been a constant nagging in our butt.
I predict that IT will adopt these tools slowly, and the breaches that resulted will permanently change our careers and financial security.
AI will not only help keep our businesses safe, but also those we care about. The aging population is the group that needs this protection most. Bad actors will often use breaches such as this to trick them into giving up retirement funds.
Only question is, will AI security be implemented before the same technology is used to harm us? AI is not good or evil, it’s just a tool. Unfortunately, new technology is used against us more often than it is for us in the field of cybersecurity.