[BreachExchange] Artificial Intelligence is the Answer to Retail’s Cyber Threats

[Audrey McNeil]

http://www.mytotalretail.com/article/artificial-
intelligence-is-the-answer-to-retails-cyber-threats/

As the cyber threat landscape continues to grow — and threat actors
increasingly steal data, commit acts of fraud, and expand retailer losses —
it’s time for the retail industry to rethink its cybersecurity strategies
and shift the focus to the protection of online outlets. As many have seen,
news of point-of-sale malware has increasingly made headlines, and these
constant threats by malicious actors are a growing concern for retailers.

Today, many retailers are leveraging artificial intelligence (AI) across
their sales operations, using the technology to help improve consumer
shopping experiences both in-store and online, determine target audiences,
personalize offers and discounts based on purchase histories, promote key
items, and more. However, many retailers are ignoring one of the most
impactful uses of AI in retail — cybersecurity.

As more people than ever before abandon brick-and-mortar stores and flock
to e-commerce sites, the potential threat to retailers from hackers is
increasing exponentially. As the volume of online transactions rises, and
fraudulent purchase attempts, denial of service (DDoS) attacks, and
detrimental bot behavior spike, retailers are at risk of losing thousands
or even millions in revenue. What’s more, these attacks could leave
retailers’ reputations in shambles and struggling to regain consumer trust,
as exemplified by organizations such as Target, Neiman Marcus and Forever
21.

These retailers are among the organizations most vulnerable to cyberattack,
given the nature and volume of customer information available to cyber
criminals. They're also some of the most advanced in leveraging new AI
technology to achieve a multitude of business objectives. So why have they
continued to overlook AI as a potential solution to their cyber woes? And
how can AI transform the future of security for retailers?

Many online retailers use off-the-shelf applications that often come
packaged with known vulnerabilities — and both researchers and hackers are
discovering numerous new vulnerabilities in these applications every day.
Other retailers may use their own customer applications, but these could
potentially expose undiscovered vulnerabilities to hackers as well.

Today, many retailers are developing their own mobile apps that use
machine-to-machine communications via APIs that can also be targeted by
hackers. Online web applications and APIs are increasingly becoming points
of entry for hackers, and must be protected from the constant increase in
cyberattacks.

For years, rule-based web application firewalls (WAFs) have been the
technology of choice to protect online retailers’ web applications.
However, these technologies have only offered generic and limited rulesets
with inadequate bot management capabilities using antiquated challenges
like CAPCHA. Those responsible for WAFs are often frustrated by the
time-to-tune, limited protection, lack of accuracy, and the increasing
management overhead. In addition, most WAFs on the market offer little if
any protection for APIs. As a result, APIs are often left completely
exposed to threat actors. However, all of this is about to change.

Today, the industry is witnessing a new generation of WAFs, bot managers
and API defenses that are incorporating machine learning (ML) and AI
capabilities into the mix. For example, forward-thinking WAF vendors are
beginning to integrate traditional rule-based WAFs with supervised ML that
can shorten the rule-tuning process to hours instead of days, weeks or even
months. ML- and AI-enabled WAFs often don't use the traditional core
rulesets, but instead identify behavioral-based anomalies and subsequently
are able to defeat a broader range of targeted application-based attacks
without the use of rules and signatures.

Beyond WAFs, these same vendors are also incorporating ML and AI
capabilities into their bot management solutions that are capable of
finding the smallest of subtleties to help identify human visitors vs.
malicious bots. Although some bots are needed to improve SEO, the vast
majority of bots have malicious intent, and their traffic targeting
retailers’ applications and APIs is not needed — nor desired. ML and AI
capabilities can also be incorporated into protection mechanisms designed
specifically for APIs, as the nature of the traffic is much different than
traditional browser-to-application traffic, whereby traditional WAF
rulesets have little value.

The recommendation for online retailers is to begin to expand their
adoption of ML and AI in the fight against cybercrime. The technology is
here, developed and ready to act as a completely independent, autonomous
system that can be deployed as a layer over human talent and traditional
WAF technology. AI is limitless and nonlinear, smarter and faster than any
human or computer, and the more it’s trained, the more powerful it becomes.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.riskbasedsecurity.com/pipermail/breachexchange/attachments/20180315/1684095c/attachment.html>