Fueled by the exponential growth in mobile messaging, chatbots — interactive messaging bots that harness recent advances in artificial intelligence and machine learning — are the hottest new technology going right now. Facebook opened up its Messenger platform to bot developers earlier this year; messaging app Telegram is offering developers up to $1 million in prizes to develop bots that are fast, useful, and work in inline mode; and over 20 million people chat with the Xiaoice bot on the Chinese micro-blogging service Weibo. Even the White House has gotten into the act with its Obama Facebook chatbot.
Chatbot technology is still in its infancy, but it’s quickly being embraced by businesses because of its vast potential for sales, marketing, and customer service. Chatbots stand to help organizations build deeper relationships with their customers and improve service quality, while at the same time save money by automating certain administrative tasks.
However, as organizations build and deploy enterprise chatbots, it’s important to step back for a moment and consider the security implications of this brave new technology.
Be Aware of the Chatbot’s Channel Encryption
For maximum security, chatbot communication should be encrypted, and chatbots should be deployed only on encrypted channels. While these sound like obvious safeguards, unfortunately, it’s not that simple. An in-house bot that runs on an organization’s system can be set up on a private, encrypted channel, but if an organization wishes to deploy a chatbot on a public channel such as Facebook Messenger, it’s at the mercy of that platform’s security capabilities.
While Facebook is testing end-to-end encryption for its Messenger platform, the feature is still in beta and isn’t widely available. Until public channels begin offering encryption services, organizations should be wary of the type of chatbots they employ using those platforms. Chatbots used on unencrypted channels shouldn’t accept or transmit sensitive information, and for the protection of the organization, these bots shouldn’t have access to the organization’s systems.
Establish Rules Regarding Chatbot Data Handling and Storage
By their nature, chatbots collect information from users; that’s how they respond to questions, and it’s how they train themselves to get better at their “jobs” over time. Where this information is stored, how long it’s stored, how it’s used, and who has access to it must be addressed, especially in highly regulated industries that handle very sensitive information, such as healthcare and finance. Before implementing a chatbot, organizations must establish rules regarding the data the bot will gather and make these rules clear to the customers who will be using the bot.
Additionally, companies must consider where this data will reside, especially if the bot collects personal or sensitive information. This is another issue that limits the functionality of bots on public platforms until the platforms can ensure secure storage and provide additional tools regarding what gets stored and for how long.
Be on the Lookout for Criminal Chatbots
Finally, organizations must be aware of the bigger picture of chatbot security. As chatbots become better at imitating humans, the technology will be used by hackers in phishing schemes and other social engineering hacks. For example, a chatbot designed to imitate a customer or a vendor could strike up a conversation with an employee through a messaging app. After rapport has been established, the chatbot could entice the employee to click on a malicious link or hand over sensitive information.
This has already happened on the consumer level; recently, a number of men using the Tinder dating app were swindled by a bot that pretended to be a female user. After a few back-and-forth messages, the chatbot convinced the men to click on a link to become “Tinder verified.” The link required that they input their credit card information, at which time they were unwittingly signed up for a recurring online porn subscription.
Until technology can be developed to identify and intercept malicious chatbots, the best defense is to train employees to never click on links sent by customers or vendors, and to prohibit them from transmitting sensitive information through email or messaging services. Organizations should be doing these things already to defend against “traditional” phishing schemes.
Because chatbot technology is so new, specific security protocols are still being developed, particularly regarding chatbots deployed on public platforms such as Facebook, and the rapid pace of chatbot development means that new features — and threats — are continuously emerging.
Chatbots have the potential to transform how business is conducted online. They can also be quite destructive and end up causing cybersecurity nightmares for organizations that don’t employ them properly, especially at this early stage. It is critical for organizations to not get caught up in the frenzy surrounding this new technology and to take a conservative, deliberate approach to chatbot development and deployment, particularly on public platforms.