As organizations increasingly adopt AI tools, it’s crucial to implement basic safety measures to help maintain your competitive advantage, prevent costly breaches, and preserve client trust. But there are so many considerations, where do you start? Here are six essential AI safety tips every leader should follow:

1. Choose Which AI Tools You Will Trust with Your Data

There are third-party tools that offer features such as recording and summarizing meeting notes, ingesting all your data to augment their responses, and more.

Review their privacy policies before you use the tools. If it states the tool and company keep your information private, but then explains they share data with third parties over whom the provider has limited control, treat the tool as having no meaningful privacy protections.

Sharing sensitive information such as your customers’ information, business practices, or anything else you want to protect, with third parties can be concerning, as it could go anywhere those third parties want to share it.

That’s why some organizations stick with the primary chatbots that are under more scrutiny. But don’t give up on the third-party tools; some of them can be very useful. Just be sure to weigh the risks of sensitive data exposure vs. the benefits.

2. Clear Your Chat Histories Periodically

Chat histories are very useful for going back and picking up conversations where you left off, potentially weeks or even months later. The reality is, even with a search function, it can be difficult to go back and find a specific chat when you have too many to look through.

The reason to remove old chats is so that a threat actor cannot read them if they break in with your login information or another way. If you don’t need the old chats, remove them.

Some chatbots state that they will remove your chats 30 days after you delete them. Because they can change frequently, always check the current policy for all tools.

Some enterprise subscriptions to chatbots permit your IT department to set policies to automatically delete all chats older than the number of days you specify.

3. Disable Automatic Sharing of Meeting Notes

Meeting notes are unreliable until a human edits and finalizes them.

If you’ve used AI at all, you’re familiar with the term hallucination. Participants in the meeting know the context of the meeting; AI must attempt to figure that out. AI tools are often designed to estimate and present the most likely meaning of conversations, even when they’re not certain.

If you have a meeting where people use a lot of words like “it,” “they,” “that,” “thing,” and so on, AI sometimes guesses what they mean, and it might get everything so wrong that the summary is inaccurate. Sometimes it can get the meaning in the notes that’s exactly opposite of what was really discussed.

A key step is to disable the automatic sharing of meeting notes after the meeting finishes. The meeting notes must always be reviewed by a human, preferably you, so you can correct any mistakes in the meeting summary before sending them out. There may be people who make decisions, important ones, based on the meeting summary. Meetings contain tasks assigned and accepted, status of decisions, and other key information, so it’s essential to confirm the accuracy of the summaries.

Some organizations have elected to completely omit recording meetings to protect the privacy of the meeting and prevent inaccurate summaries from leaving their organization. If they do have AI make notes, they think twice before sending them to someone outside the organization. If meeting notes or a summary contain misinformation that leaks, you have no control of information already sent.

4. Anonymize Member or Client Information When You Give Information to AI

For example, if you’re creating a sensitive email to someone who’s upset, you might substitute a fictitious name for the person’s real name and the organization’s name, just in case there’s an information leak. Anonymization can be very simple: just use the word “Jim” where you would normally use “Tom.” This one’s up to you, but some people sleep better at night knowing they didn’t put their customer’s actual name into the AI tool.

Then, after you finish tuning up your correspondence, before you send out that message or that document, you simply do a find-and-replace to restore the names of the person and the company to their correct names. And you’re doing that outside of the AI tool.

Many people forgo anonymization most of the time because it adds two extra steps, but they use it in special cases. Keep in mind that changing people’s and organizations’ names might still not be enough to anonymize the discussion if you enter a unique event, location, project name, or another bit of context that ties back to the actual person or organization.

5. Disable the AI Model’s Training Features in the Settings

The most common concern I hear from business executives is that their organization’s sensitive information will leak into the public domain. The term “training” describes a large language model learning from your chats. If you provide information such as a customer list and the training or learning is disabled, the chatbot should not remember your sensitive information or share it with another user at another company, unbeknownst to you, anywhere on the planet.

Most chatbots allow you to disable learning or training based on the information you enter, and sometimes the training setting is “off” by default.

Disabling training typically means your data is not used to improve the public AI model. There is no guarantee that data isn’t stored, reviewed by a human, or exposed through a security incident.

6. Always Use Strong Passwords and Multi-Factor Authentication on All of Your AI Accounts

If a stranger or other unauthorized party were able to log in to your chatbot account, they could read all your saved chats and learn a lot about you and your organization. They can craft fraudulent email messages so accurately that you or members of your team would fall for them without hesitation. Threat actors could also use your chatbot in unethical ways that would appear to be you. You could get locked out of your account for misbehavior. Another risk is that threat actors are designing tailored prompts that cause chatbots to bypass their alignment boundaries. Furthermore, attackers can use compromised chatbot accounts as a trusted pathway into systems and data. Just as you benefit from AI’s power, the attackers can use your AI’s power against you.

As with any website or service, use the strongest sign-in protection the chatbot supports. Using a password alone is considered insufficient authentication protection. Passwordless multi-factor authentication is usually the strongest option available and relies on your phone, fingerprint, facial recognition, a physical USB key, or another method that doesn’t require entering a password but still has more than one factor.

If the login doesn’t support passwordless login, using an authenticator app on your phone with number matching is sometimes the next best option.

If an authenticator is not available, use a text or email message as your second factor. It is far better than having no multi-factor authentication.

Always remember that authentication protection, no matter how advanced, is not immune to threat actors using techniques to bypass MFA. Always be wary of unexpected login prompts, as they may be attempts by a threat actor to gain access through you.

Conclusion

Those are some basic AI safety tips for leaders. These are all very simple to accomplish, and there’s a good chance you’re already doing most or all of them. Please forward this to your friends so that they can make sure they’re following these steps too.

About the Author

Mike Foster, CISSP®, CISA®
Cybersecurity Consultant and Keynote Speaker
📞 805-637-7039
📧 mike@fosterinstitute.com
🌐 www.fosterinstitute.com

Mike Foster is a leading cybersecurity consultant with decades of experience helping organizations across North America secure their digital assets. He holds CISSP® and CISA® certifications and is the author of The Secure CEO. As the founder of The Foster Institute, Michael has delivered over 1,500 keynote presentations and consulting engagements, equipping executives and IT leaders to strengthen their cybersecurity posture and defend against evolving threats.

The post Six Essential AI Safety Practices for Leaders appeared first on Foster Institute.

Logins often go like this:
1. User enters a password
2. The user receives a text message with a code to confirm it is them

That second step is called MFA (Multi-Factor Authentication).

Entering numbers is frustrating, so some organizations use push notifications:
1. User enters a password
2. The user has an authenticator app asking, “Is that you trying to log in?” The user can click “Yes,” and they’re in. If it is an attacker, the user can ignore the alert or click “no.”

Attackers are overwhelmingly bypassing push notification protection. Here’s how: Once the bad guys obtain the user’s password, they make frequent attempts to log in, sending many alerts to the targeted user’s phone. The user is annoyed by the interruptions. Sometimes, users click “approve” to stop the annoyance. They don’t realize they are allowing an attacker into the system. This phenomenon is known as MFA fatigue.

If you use push notifications, ensure that your push notifications require the user to enter at least two digits displayed on the device where they entered the password rather than simply pressing the approve button in their phone app. This verification helps prove the user has the device. The process becomes:
1. The user enters a password on the device they’re logging into.
2. The user has an authenticator app asking them, “What code is your device displaying?” The user enters the code, and they’re logged in.

Microsoft calls the process “Number Matching.” Duo calls it “Verified Duo Push.” Okta calls this a number challenge.

If you use push notifications, enable some form of number matching verification ASAP to combat MFA Fatigue.

Please tell your friends to be sure they know this hidden danger and can tighten security in their organizations.

The post MFA Fatigue: The Hidden Danger and How to Combat It appeared first on Foster Institute.

Here is how the scam works: Suppose you want to look up a company online named Super Duper, so you type the store’s name into your favorite search engine. An attacker might have purchased the top result to take you to the website superduperco.com. However, if you knew to scroll down past the paid-for-results, you would have seen that the real website is superduper.com. Attackers set up a website and named it superduperco.com.

Their deceptive site might contain malicious advertising, ask you to enter credit card numbers during checkout, or tempt you to download malicious programs and apps. They might ask you to login or reset a password, and they capture the password you type in.

If you look up a retailer in a search engine, skip past the ads and paid results. Scroll down to see real search results. Even then, be skeptical in case attackers used SEO techniques to appear at the top of the actual search results.

Please forward this to your friends to alert their users that top search engine results can be a trap.

The post Beware: Attackers Buy Top Search Engine Results to Trick You appeared first on Foster Institute.

USB: A new ploy is bad actors mailing USB devices, appearing to be from your company to your users. Once plugged in, they can open up a channel that permits unauthorized remote control and capturing keystrokes, including passwords.

Fake Login Prompts: Remind your users to beware of login screens when they don’t expect them. Attackers create persuasive prompts that ask your users passwords for their logon, VPN, or Microsoft Office 365 login, and more. Sometimes the windows tell the user that their connection dropped, and to provide credentials to reestablish their link. Your Users must ignore those prompts and notify your IT team immediately.

Fraudulent websites and apps: Sites may have useful coronavirus information, but they also contain malicious attack software that strives to infect computers. Attackers create bad apps offering online statistics, tracking of the virus spread, and more.

Please forward this to your friends so they can alert their users too.

The post Alert Your Team – USB Devices, Login Prompts, and Apps appeared first on Foster Institute.

If you want to find out if your passwords were released, visit his site called https://haveibeenpwned.com. If you elect to enter your email address, he will tell you if it is in the collection and give you more details.

What do you do if you are on the list? Reset your passwords. Use a password manager that will remember your passwords for you to make your life easier when you use a different password at each website from now on.

Now is a great time to enable two-step verification. A basic form of two-step verification is when you enter a username and password, and you receive a text message code to type in. Enable two-step verification on PayPal, LinkedIn, Dropbox, Facebook and every other web service you use. On each website, look for Settings > Security. You may need to dig down, but more reputable sites now support two-step verification, but you must enable the feature.

Some bad news is that, about a week ago, a tool called Modlishka shows how to break two-step verification so it isn’t that secure, but two-step verification is still more secure than a simple username password combination. If it allows, have a website use some other method than texting you a password. Using an app on your phone or calling you via a voice call are options that are often more secure than the text message. Microsoft, Google, and a service called Duo offer these options and more. Having a hardware key is even better unless your laptop users leave the key stored in the laptop case, and their password written on the bottom of the laptop.

The post 773 Million Passwords Exposed – Were You Exposed? appeared first on Foster Institute.

You assign your IT team the responsibility to secure and manage the keys to the vehicles, so you give each member of your IT team a copy of the master key.

Here is where it gets crazy: Suppose that there is a well-known tradition, in all companies, for IT professionals to store their master keys in the top drawer of their desks. Unfortunately, if someone wants to steal a vehicle, they know right where to find a master key. They can take all the cars once they gain access to the master, and they know exactly where to find it.

In the real world, your IT team has the responsibility to secure and manage your most sensitive data. In doing so, they have the master keys that unlock all the other keys. It is a tradition to give all IT professionals, and even outside consultants, keys to the master lockbox. The shocking part is that all IT professionals are encouraged to store the master keys in the same place, in the default well-known security groups named schema, enterprise, and domain admins.

Your IT team must create new security groups, with different names, in which to store the master keys. It is crucial that the new groups only provide specific privileges to member users on a need to know basis. It is ok if this strategy is new to them.

To measure this, ask your IT professionals to show you what users are members of those default security groups. Discuss moving those users into specific groups that provide the least amount of access they need to perform their work. Depending on the complexity of your system, this may take more time. IT professionals are always busy, so discuss with them their current projects, then prioritize this essential security improvement accordingly.

Storing master keys in a well-known location is absurd, and it is likely that you are doing that now.

The post The Insanity of Your Network – Storing Keys in the Same Place as Everyone Else appeared first on Foster Institute.

When you want privacy, one way is to encrypt your documents before you attach them to your email message.

Microsoft Office, for Windows and Mac, has a feature on the File menu called Protect Document. Choose that option, and enter a secret password.

Use a phrase such as: the chairs are in a row.

E-mail that file to your recipient.

Then, phone, or text, the password to your recipient. If you email the recipient the password, even if it is in a separate email message, whoever is reading your email messages will receive both the attachment and the secret code.

Please forward this to any of your friends who may want to send sensitive email attachments.

The post How to E-Mail Encrypted Attachments appeared first on Foster Institute.

Using plain letters is enough; numbers and symbols are optional.

Passwords need to be long, at least 8 characters. 16 or more is better.

You don’t have to reset passwords unless you suspect a data breach. Wow!

However, passwords cannot match a database of stolen or easy to guess passwords. And, your login mechanism has to have a way to verify that.

That’s the basic gist. You can read the details in the NIST Special Publication 800-63 at https://pages.nist dot gov/800-63-3/sp800-63-3.html It covers secret questions, two step login, etc.

Now, hopefully PCI and other standards will update their rules too.

Please forward this to anyone who is sick of complicated passwords and having to change them every 90 days.

The post Make Life Easier – Go Back to Easy Passwords appeared first on Foster Institute.

The most important thing about the OneLogin breach: It affects you and everyone else, not just the 2000 customers of OneLogin.

If you store information in the cloud, including information your customers entrust to you, and if your cloud provider uses OneLogin internally, then your sensitive information could possibly be accessible as well. Cloud based service providers you use every day might use identity management.

This is another example of how someone else’s breach can hurt you, including exposing your customers’ sensitive information.

The big question is: How long have attackers had access? Thank goodness OneLogin at least identified that they’d been breached. Are any other identity management firms breached and don’t yet realize it? What are attackers accessing around the world using stolen passwords?

Please forward this to anyone you know who may not realize that these single points of failure, holding login information for many services that even your service providers may use internally, are very attractive targets for attackers. One successful attack results in a goldmine of information, including yours.

The post Online Password Warehouse, OneLogin, Discovers Massive Breach appeared first on Foster Institute.

Password managers are very helpful since they make it so convenient to be secure, and can greatly simplify and speed up the login process at websites. Many people feel password managers are worth the risks, especially when the risks can be minimized as summarized below:

First, as you can see, there is no guarantee that password managers are perfect. Never store super-sensitive passwords into your password manager. Store them in your head.

Second, enable two-step verification on all websites. Then, if an unauthorized person obtains your password, they will have a difficult time logging in, if they cannot perform the second step.

Third, one of the ways to launch the exploit involves tricking the user into clicking a link, such as a link in an email message, or getting a script to run on a web page as the user visits the page. Using click-to-play can greatly minimize those risks.

To learn more about the first two, see last week’s newsletter posted at www.fosterinstitute dot com/blog/your-iphone-and-ipad-are-in-danger. Never mind the title; the content addresses the first two steps listed above even if you use Windows or Android.

As for the third point, we’ll cover click-to-play next week, or you can simply google those terms and get started right away.

The announcement came from LastPass, and don’t panic if you use it. LastPass says the exploit is very difficult for an attacker to use, but not impossible. Resetting your passwords is not going to help, yet. Only after LastPass develops a patch, and then only when LastPass on your computers are patched. LastPass said this only affects users using the LastPass extension in Chrome, but that researchers have used the exploit in other browsers too. Email us if you want more technical details.

Please forward this to anyone you know who may use a password manager or lets their browsers remember their passwords.

The post Alert – A Popular Password Manager Has Serious Security Flaw Right Now appeared first on Foster Institute.