Learn why security needs to be integrated into every AI solution to ensure a smooth digital transformation thatโs safe from vicious cyberattacks on all fronts.
After spending years as a data scientist and as part of data and analytics teams within HPE, I'm happy to see data and
AI gets realโand what that really means
AI is perhaps the most used or even over-used term of the decade. The notion of AI has been a theoretical concept (and a fascinating movie theme) for many years. Now the term has gained even more traction, thanks to the innovations and solutions in the areas of data creation, data storage, and most importantly, the use of computational power to get insights from stored big data.
Many organizations are already working toward enabling AI-driven digital transformations that cover the entire solution lifecycle from proof-of-concept to implementation. According to the International Data Corporation (IDC) Worldwide Artificial Intelligence Spending Guide, spending on AI systems will accelerate over the next several years as organizations deploy artificial intelligence as part of their digital transformation efforts and to remain competitive in the digital economy. The compound annual growth rate (CAGR) for the 2019-2024 period will be 20.1%.
Every business sector is involved in the AI adoption race. All are building innovative AI solutions to solve various business problems and to optimize operational efficiency.
With ubiquity comes greater security threats
The AI systems that organizations deploy today connect to all the key data sources, consume their most critical data sets, and make mission-critical decisions.. Because of their important role in the monitoring of organizationsโ operations, these AI-driven devices have become a target for cyberattacks. Certainly, hacking or compromising these critical devices could have a significant impact on organizationsโ ecosystem. That is why vulnerabilities left open in AI systems are more dangerous than on other systems.
Attackers act much faster than we doโand business leaders should be ready to address this. A good example of this is the attack of the Microsoft Tay Chatbot that was tweeting racist comments after attackers poisoned its learning algorithm. The use of chatbots in customer service industry is growing. According to estimates, more than 67% of consumers worldwide used a chatbot for customer support in the past year and around 85% of all customer interactions will be handled without a human agent by 2020. All make suitable targets for data poisoning and intent manipulation attacks.
Another example is McAfee's hack on older Tesla MobileEye cameras, which tricked the cameras into misreading the speed limit. A recent article by MIT pointed out an attack vector on AI machines whereby systems would be forced to consume more energy which would indirectly lead to a denial-of-service.
Recent trends in AI attacks target machine learning models and libraries, including state-of-the-art neural networks that are vulnerable to cyberattacks, which can in turn lead to adversarial behaviors.
Why security is needed at every step of AI solution development
All organizations must be aware of these potential threats. The impact can be tremendous if security is not integrated at each stage of the AI solution development lifecycle. Published in 2019, Gartnerโs predicted that โThrough 2022, 30% of all AI cyberattacks will leverage training data poisoning, AI model theft, or adversarial samples to attack AI-powered systems.
Today, the AI attack surface is also expanding right along with the increase in AI adoption. AI has become a key pillar of every digital transformation and security is a key component of AI. Security controls can protect machine intelligence and drive a trusted adoption of AI.T
To sum it up: The building of secured AI machines is key for the success of most organizationโs digital transformation.
How to protect smart machines from attacks?
Until recently, cybersecurity was primarily reactive security, meaning it focused on securing the IT infrastructure and then responding to threats. Now the focus is shifting to the creation of proactive and predictive security controls that use AI.
Software is prone to being hacked and infected with computer viruses, or its users might be targeted by scammers using phishing and other security-breaching ploys. AI-powered apps are no exception. Securing intelligent machines requires specially designed security controls along with current cyber protection measures. It also requires additional layers of security to cover key blocks of AI architectureโwith a deeper focus to control adversarial attacks along with traditional attacks.
Significant growth in the number of new attacks enabled by AI, instances of new malware, and varieties of existing malware justify the need for AI in defending the security systems. These systems can process, detect, identify, and remediate many types of threats in microseconds. Machine learning and deep learning models are currently used in security for risk sensing, threat modeling, and identification and monitoring along with risk process automation in the IT security space for predictive risk intelligence.
Each of these methodologies offers benefits for addressing specific security problems. Merging all these solutions can secure data and models while also controlling unethical behavior to enable risk-aware decision-making. As a result, organizations can build AI-driven defenses to better prevent attacks and protect data and assets even from adversarial attacks.
The data and insights AI security solutions provide can become the ultimate security intelligence source which can strengthen evolving AI machines and turn them into secure and trustworthy AI systems.
How HPE can help
The right AI solutions
Start unlocking the value of your data with innovative, flexible AI solutions from HPE that are designed to give you the scalability, performance, and cost controls you need.
We make AI that is data-driven, production-oriented, and cloud-enabled, available anytime, anywhere and at any scaleโand always with security front of mind. Our solutions support AI for the enterprise, public sector, financial services, healthcare, life sciences manufacturing, and more.
Learn more: HPE artificial intelligence solutions
The right mix of AI security skills and experts
Wherever you are within your AI journey, itโs never too early to start thinking about security, risk, and compliance requirements. Advisory and professional services experts with HPE Pointnext Services already work with number of organizations to assess business needs. We help architect, design, and implement a secure AI framework by integrating security controls at every stage of an AI solutionโfrom edge to cloud. Our experts have many years of experience in building and implementing complex security solutions for a wide range of problems across industries and around the world. Our team also partners with leading security solution vendors to protect data, platforms, and data insights as part of our AI security offerings.
As a best practice, HPE experts combine AI, data, cloud, and security expertise to build security-embedded data platform solution reference architectures that are specially designed to protect AI implementations from attacksโespecially adversarial attacks. Our framework is also aligned with the NIST and ISO AI security standards and
policies, as well as the MITRE-proposed threat matrix.
Learn more: HPE security and digital protection services and HPE AI and data transformation services
Please contact us if youโd like to discuss further.
Related articles:
- Security: The foundation for transformation success via HPE's Enterprise.nxt
- The rise of artificial intelligence and machine learning via HPE's Enterprise.nxt
- Stay current on top tech trends and expert advice. Sign up for the weekly newsletter.
Rohini Chavakula
Hewlett Packard Enterprise
twitter.com/HPE_AI
linkedin.com/showcase/hpe-ai/
hpe.com/us/en/solutions/artificial-intelligence.html
R_Chavakula
Rohini is a data scientist in HPE GreenLake Cloud Services where she works on building trustworthy AI machines. Rohini advises and designs responsible AI systems for trusted outcomes. Working with the security practice and building AI solutions to tackle business challenges across domains have combined to foster her interest in AI security.
