As the adoption of artificial intelligence expands, it offers cybersecurity leaders multiple benefits. They just have to focus on tangible results.
Federal agencies are adopting artificial intelligence to improve cybersecurity. These technologies could one day be fully automated to protect networks and scale defenses more quickly from incoming attacks.
AI will dramatically change the way we protect networks by automating certain tasks that free employees to focus on larger security goals. Fully autonomous systems, of course, are far away, but the steps agencies take today may lead them more quickly to that destination.
As agencies continue to implement more ambitious AI and machine learning programs, they will see improved functionality, more efficient systems, and lower costs, among other benefits. The key for agencies today will be to continue to adopt ambitious AI solutions that deliver tangible results that can be clearly shown to stakeholders.
From Narrow AI to Strong AI
Right now, agencies have moved slightly past an important stage called narrow AI where artificial intelligence technologies require custom processing, usually through purpose-built
on-premise appliances or cloud services. These tools also require specific training of rules-based models that run algorithms to handle specific tasks but need human oversight.
Many agencies have reached or moved past this stage. They have implemented practical applications that help with specific tasks.
This is the case at the Homeland Security Department’s Cybersecurity and Infrastructure Security Agency (CISA), where Martin Stanley led the development of selection criteria for AI projects with “high benefits and low regret.” He wanted projects that could show stakeholders – from program managers to agency executives – the advantages AI presents. This criteria also emphasizes projects that would not cause harm if the AI implementation somehow failed.
Stanley, now the senior technical advisor for the Office of the CTO at CISA, shared this information on a recent webcast with me.
Projects like the one at CISA will help agencies move closer toward strong AI, also called true AI. At this stage we begin to see the true benefits AI provides. True AI features purpose-built models deployed across the enterprise and accessible on-edge devices. This form is accessible to end users and takes advantage of deep learning and cloud computing to run unsupervised methods of the technology.
The Future of Cyber
The future of AI in cybersecurity means a world where humans collaborate with machines and redefine business processes, one in which people are still at the center. Some of the possibilities for the future of cyber include:
- Security orchestration that enables a coordinated response to incidents using all available tools.
- Automated response where AI responds in near zero time with coordinated tools.
- Cognitive capabilities where the AI uses an “intuitive leap” to create truly teachable software based on the current threat landscape.
- Automated analytics where machine learning algorithms draw connections beyond what is in the data.
A Technology in Demand
While some of these future goals for AI in cyber seem far off, they may come sooner than many expect based on increasing interest from technology leaders. In Accenture’s Third Annual State of Cyber Resilience Report, which surveyed 4,644 security executives from 24 industries in 16 countries, including 100 from the federal government – AI was named either the first or second most important technology in addressing certain areas of cyber resilience. They include reducing breach impact, ensuring consistent quality of response, more precise incident detection and cost reduction.
It is no wonder that the market for AI in security is expected to grow with a 35 percent compound annual growth rate over the next four years.
AI continues to transform how we protect networks and data. It is already providing value to leading government agencies, and we will continue to see advances in perimeter and endpoint defense, digital identification, security information and event management.
This will only improve. Historically, the government procurement process has languished behind the pace of commercial innovation, but this is changing. Thanks to the NASA Solutions for Enterprise-Wide Procurement (SEWP) V governmentwide acquisition (GWAC) contract vehicle, all federal agencies can more speedily gain access to mission-critical cloud and cyber technologies and services, including Accenture’s Extended Detection and Response (XDR) for government, FedRAMP authorized platform, which is available through reseller Federal Resources Corporation (FRC).
The NASA SEWP V vehicle provides cybersecurity, managed services, platform as a service and continuous innovation to government agencies, and also includes innovative AI technologies to boost cybersecurity. As more agencies use AI and shared services to protect networks, it will reduce the daily burden on security staff, allowing executives to focus attention and resources on larger security goals and other technology initiatives. One day AI will provide a dynamic response to cyber-attacks that can outperform the most seasoned security professional.
The federal government’s path with AI is seemingly just starting with the best innovations still yet to come. Security leaders today must embrace AI, but also communicate its benefits to non-technical stakeholders within their ecosystem. AI is the future, and even after more than 16 years in the security field, I am excited to see where it can ultimately take us.