A.I. Platform FAQs and Best Practices

by | Sep 9, 2023 | Business, Marketing, Observations, Operations, Safety, Security, Social Media

AI Image, man holding a digital lightbulbThis is my 2 cents based on what I know about AI so far. Since I am still spending several hours per day learning new AI platforms, and new information it’s going to be a moving target info wise, this is my knowledge and my personal take on all of this so far.

 

AI Platforms output is not always correct, ALWAYS fact check especially if you are using it for a research tool. Hopefully that information is out there already for most people, but I’m still running into people just testing the waters for the first time that don’t actually know that.


Keep in mind anything added to AI engines is being used to train the AIs; don’t ever put confidential or private information in. Also, be cognizant of putting any personal information in. Just like you shouldn’t be using your favorite vacation spot as a password, hackers are also using personal information to train AIs to hack into accounts.                 

 

There has been a recent uptick in hacked social media accounts that is being attributed to AIs being trained by hackers to browse for personal information used publicly online, and that can include data being put into AI generators.

Some AIs are allowing you to opt out of data gathering as well as some platforms and programs that also integrate AI into their software.

 

To opt out of ChatGPT

  • Go to your account at far left bottom, click on the three dots
  • Go to settings
  • Go to the Data Controls tab on left
  • Slide “Chat history & training” to off

 

Meta/Facebook

 

Adobe

 

Don’t use text straight from an AI platform for website text, blog text, or any other online published article text (social media posts are an exception as far as I can tell). Always rewrite and reword a bit. Google can tell it’s AI-generated content, and it can affect your Search Engine Optimization (i.e., how a business website gets found in Google search)


Don’t sign into any AI platform with your Google account or social media accounts, ALWAYS use an email and a unique password to login. Giving access to your Google account creates a backdoor which from a business and personal standpoint, can compromise your email account, and any other Google accounts you own: (Google Suite, Google Drive, Google My Business, Youtube, Google Docs, etc.). 

 

If you are not familiar with Data Breaches and the cost of how it can affect a business (The global average cost per data breach was 4.45 million U.S. dollars in 2023.), please read https://www.upguard.com/blog/cost-of-data-breach and https://www.statista.com/statistics/273575/us-average-cost-incurred-by-a-data-breach/ . 

 

This is also how many social media accounts get hacked. Granting access to Facebook through a backdoor can cause an account to get hacked, and just changing the password to the account, does not remove access, you have to remove app access too.)

 

If someone does give access to a Google account, and needs to unlink it/revoke permissions , here is a walkthrough https://support.google.com/accounts/answer/2541991?hl=en#:~:text=Select%20Connected%20accounts.

 

Revoke app access/external Access (social media)

Facebook https://www.facebook.com/help/942196655898243 

Instagram  https://help.instagram.com/588549329146493 

Linkedin   https://www.linkedin.com/help/linkedin/answer/a519947 

Pinterest   https://help.pinterest.com/en/article/connect-to-other-apps-with-pinterest 

X (formerly Twitter) https://help.twitter.com/en/managing-your-account/connect-or-revoke-access-to-third-party-apps 

Tiktok https://support.tiktok.com/en/safety-hc/account-and-user-safety/connect-to-third-party-apps   

 

Please be extremely careful adding browser extensions to your browser to extend AI capabilities, there are A LOT of unsafe, unsecure and malicious browser extensions out there that can compromise people’s computers and contain malware or keyloggers, a good overview of browser security can be found here https://security.berkeley.edu/education-awareness/browser-extensions-how-vet-and-install-safely .

 

Malware can use known software vulnerabilities to infect your PC. A vulnerability is like a hole in your software that can give malware access to your PC. When you go to a website, it can try to use vulnerabilities in your web browser to infect your PC with malware. Apple computers can also be compromised as malware and virus programs designed specifically to attack Macintosh operating systems can be even more malicious and destructive than ones designed to attack PCs.

 

A keylogger is a form of malware or hardware that keeps track of and records your keystrokes as you type. It takes the information and sends it to a hacker.

 

A best practice is to NOT download any AI software. If there is a web-based version, use that instead. Before downloading any kind of software, do your research. Just Googling the name and reviews, or the name and whether it is safe will not give you legitimate search results. Insecure and malicious sites and software have a lot of sites that are either set up by them or paid to have them set up to make software come across as safe and legitimate. If you do download anything, make sure you have a good antivirus AND a good malware program, and it’s up to date; scan the program files before installing. When in doubt (even a little) don’t do it.

*Most antivirus programs say they cover malware, but they only cover a fraction of what’s out there. It’s best practice to have both. On average, over 500K malware and virus programs are created daily. I recommend Malwarebytes (free and paid versions are available). There are other good ones out there as well. 

 

Be cautious when using AI photo editors/generators. This is a big gray area online at the moment, but watermarks and digital watermarks are being pulled into AI-generated works. Remember that AI-generated art aggregates other online media and creates something with it, including copyrighted artwork and photos. 

 

Currently, it’s best practice not to use AI-generated art on anything online “owned” by a business—a website, blog, etc. There is a spate of lawsuits going on currently regarding copyright, and until things are officially ruled on, it’s better not to take a chance. I look at it this way: if a business gets sued for using a copyrighted image that was integrated and generated by an AI platform, the AI platform has Terms of Service that give themselves a waiver (the platforms) from liability but not the business, and even if the suit does get dismissed, there is the time, money and reputation of the business that is risked. 

 

Until the courts have consistently ruled about this and more AI art generators start to accept their own liability for using copyrighted art in generated work, it’s better to be safe than sorry. 

 

If a business gets hit with a claim by Getty images, the beginning claim will start at around $1500, they don’t care if you remove the image online, they WILL send collections to harass a business until they pay up. If you want some background on Getty copyright claims just Google: “Getty images copyright claims small businesses” or email me. I’ve been told horror stories by hundreds of businesses, and had to help several dozen businesses over the years who have been hit with Getty claims.

 

Keep in mind as well that many AI art generators claim copyright to the images produced and many only state that you can use the generated images for non-commercial use.

 

(September 7, 2023) Microsoft announced it will protect users from Copyright claims. (Which is terrific news, hoping additonal platforms will step up to this).

https://blogs.microsoft.com/on-the-issues/2023/09/07/copilot-copyright-commitment-ai-legal-concerns/ 

 

Canva (Canva.com) terms of service regarding AI generated content (as of 9/8/2023)

https://www.canva.com/policies/text-to-image-terms-2022-12-06/    *Note: Read the Legal Notice

I’ve talked to Canva about this, and they refer to this (above link) and have stated by email that to date, they will not be liable for any copyright claims. I first found this as a potential issue because certain prompts in their text to AI image generator, particularly using the watercolor feature, return a large portion of the time images that look like they have copyright signatures in them.  From our 17 email exchanges about this topic: “We understand that the appearance of signatures on AI-generated images can be surprising and even concerning. We’re exploring solutions to mitigate their occurrence. However, the signature generated by Text to Image is just that: generated by the AI model. The AI model (Stable Diffusion) has learned from reviewing billions of pieces of data that some forms of images, such as paintings, often have a signature in a bottom corner, so it generated its own variation of one in response to the prompt.” Interesting that the AI platform only seems to be assigning “create a watermark” to the watercolor feature and not concept art, color pencil or ink print, to name a few options, and it doesn’t explain things that pop up like on the image to right where you have a double signature…….  I have noticed (so maybe they paid a little attention) that the number of images has decreased in recent testing, but the one on the right I generated on Friday in Canva, so they still have some work to do. And since Getty Images is in the midst of a lawsuit with Stable Diffusion about this very topic…….

Example:

I love Canva, it’s incredibly useful, don’t get me wrong, but I also have a much greater concern that someone i.e. a small business, will use an image generated by the Canva Image generator and get sued not understanding that Canva isn’t going to protect them. So make sure the TOS is read and more importantly understood!

 

Stability AI  https://stability.ai/terms-of-use   (as of 9/8/2023)

Midjourney  https://docs.midjourney.com/docs/terms-of-service   (as of 9/8/2023)

DeviantArt  https://stability.ai/terms-of-use  (as of 9/8/2023)

 

The better the prompt the better the results, a good resource for learning how to use prompting is https://learnprompting.org/ 

 

I found using several AI platforms at once and putting in the same prompts gives in some cases, very different and varied feedback/results, if you’re using it for marketing, use it to cherry pick the best of the best.

 

Useful AI Platforms (there are many many more)

https://chat.openai.com/  (free and paid versions, the paid version is $20 per month)

https://claude.ai/chats

https://bard.google.com/ 

https://www.bing.com/?scope=web&cc=US 

https://elicit.org/ (for research from papers and publications) 

 

How can it help small businesses? and What can AI be used for?

 

To date this list (and it’s the tip of the iceberg) is what “I” have used AI platforms/generators/sites for and found useful. There are 100’s of additional uses besides these listed, these are just the ones I’ve used so far.  I find I’m using AI on a daily basis as it is so incredibly useful for a wide variety of things.

 

For marketing and business development:

  • Research (always fact check)
  • Business and consumer trends (always fact check)
  • Article rewrites  (see above #3)
  • Social Media Posts
  • Email scripts and scripts for autoresponders
  • Keyword Research (always fact check)
  • Generate video scripts
  • Creating chatbots for customer service
  • Automating emails
  • Writing social media ad copy
  • Writing print ad copy
  • Writing Google Ad words and Bing Ad words copy
  • Create, edit, and modify media and video files (mostly paid AI platforms including ChatGPT plus)
  • Blog titles
  • Blog posts  (see above #3)
  • Email marketing subject lines
  • Email marketing content
  • Website content* (see above #3)

 

For operations:

  • Training Manuals
  • Employee Manuals (always run by an HR pro)
  • Transcribing text
  • Automating tasks
  • Voice assistants
  • Help with grant writing/grant writing
  • Cover letter and resume writing
  • Inexpensive marketing ideas for small businesses
  • Generate customer surveys
  • Proofreading and Editing
  • Create contracts (always run by a business lawyer)
  • Scripts for lectures and talks (or a speech for an event or wedding)
  • Lesson plan outlines
  • Writing and debugging code
  • Creating WordPress plugins
  • Creating WordPress themes
  • Write product descriptions
  • Create tests
  • Explain complex topics in layman’s terms
  • Create content in other languages (please use a human being to double check)
  • Create Interview questions
  • Translate text
  • Create outlines for a book or other content
  • Creating job descriptions
  • Creating job applications
  • Creating job search listing content/text
  • Extract data from text
  • Product/service name ideas (same caveat as business name ideas below/Please check!)
  • Business name ideas** (see below screenshot). Please make sure any business names are checked online before registering, in testing most of the major AI generators/platforms they are ALL making suggestions that when do a search in Google, many of the business names are already being used. This is a screenshot from the paid version of ChatGPT. Highlighted names for example are names already in use (9/8/2023)

AI Generator Business names

Again this is the TIP of the iceberg on what a business can do with AI, with the exception of the video editing, all of the above were done using the free versions and free tools that were out there without adding browser extensions, or giving access to my Google account. So until the wild west of AI platforms has some more security and rules and regulations in place, how safe do you want to keep your business??

 

AI Program/Platform Directories (Many more out there, I found these useful to start)

 

One of the good Newsletters I subscribe to: (there are hundreds +++ out there. This one is an aggregate)

 

Articles of Interest:

How Your Facebook Business Page Got Hacked

by | Sep 27, 2022 | Business, Facebook, Operations, Safety, Security, Social Media

You Have Been Hacked Image

1. You allowed unsecure or compromised apps (applications) access to your personal Facebook page (and, through it, your Business page). These can be old, outdated Business page add-ons that are no longer supported by the developers or by Facebook, or they can be malicious applications that deliberately want to gain access to your account.

 

2. You had an employee or used an external company, and you gave them full access to your Facebook business page. You had a falling out with them or fired them, and they kicked you off your Page as an administrator and then compromised your Page or deleted the Page.

 

Reason One (more information) and what to avoid.

It would be best if you give as few external applications as possible access to your Facebook account (and, through that, your Business page). Most people grant app access in a variety of ways, to play Facebook games or to take silly quizzes (please do NOT do this!) as many of these are purposefully malicious or they are easily compromised, and though that can compromise your account.

 

The other way to be compromised is to log in to external sites using your Facebook account. Zoom, for example, gives you the option to log in with Facebook; it also gives you the option to log in with your Google account (please DON’T do this either, it’s even worse). Always create a new account using your email address and a unique password, don’t ever connect social or Google apps. Some newspapers and blogs offer social sign-ins and Google sign-in options as well; please create a unique account with these platforms. An external account gets hacked or compromised, so can any accounts it has access to from the back end.

 

If your account was compromised through an external application, changing your password in Facebook does not stop the hack; you need to remove the app at fault (I always recommend disconnecting everything and readdressing and reconnecting if there is one you need later on) and then change your password.

 

See the link at the end for how to access where your app access is located so you can check and disable them.

 

Reason Two (more information) and what to avoid.

 

A Facebook business page owner should be the only person with “full” access to the Page and at least one additional person who you implicitly trust: a spouse, a business partner, or your BFF from grade school.

 

Why have that? What happens if your personal account is hacked or you get locked out of Facebook, either temporarily or long-term? Facebook has no support system per se, no phone number to call for help, and the only way to “sometimes” get them to respond is multiple support requests, and I mean dozens of them. Sometimes…….

 

Giving an employee or manager, or external company full access to your Facebook page can be a recipe for disaster. If you do, please make sure you trust them! I could tell you countless horror stories about employees or managers being fired and still having access to a page. You can guess what happens from there. I know one marketing company in New England that had a tiff with a tourism group about seven years ago, and they deleted the tourism group’s Facebook page, of which there were over 6000 followers on the Page. Facebook will not restore a deleted page if you don’t have access to it as an administrator. And now, with the New Page Experience, it looks like it can’t be restored at all.

 

Anyone helping with a page can get more limited access (which means they can’t delete the Page or add or remove admins).

 

Currently, most pages still have multiple levels of access:

  • Admin: Can manage all aspects of the Page. They can publish and send Messenger messages as the Page, respond to and delete comments on the Page, post from Instagram to Facebook, create ads, see who created a post or comment, view insights, and assign Page roles. If an Instagram account is connected to the Page, they can respond to and delete comments, send Direct messages, sync business contact info and create ads. This person can manage everything you can, including the ability to give access to others, remove anyone from the Page (including you) or delete the Page.
  • Editor: Can publish content and send Messenger messages as the Page, respond to and delete comments on the Page, create ads, see who created a post or comment, post from Instagram to Facebook, and view insights. If an Instagram account is connected to the Page, they can respond to and delete comments, send Direct messages, sync business contact info and create ads.
  • Moderator: Can send Messenger messages as the Page, respond to and delete comments on the Page, create ads, see who created a post or comment, and view insights. If an Instagram account is connected to the Page, they can respond to Instagram comments, send Direct messages and create ads.
  • Advertiser: Can create ads, see who created a post or comment, and view insights. If an Instagram account is connected to the Page, they can create ads.
  • Analyst: Can see which admin created a post or comment and view insights (i.e., statistics).

 

The “NEW” Facebook Page Experience:

People with Facebook access

  • Content-Create, manage or delete posts, stories, and more as the Page.
  • Messages-Send and respond to messages as the Page.
  • Community Activity-Review and respond to comments, remove unwanted comments and report activity.
  • Ads-Create, manage and delete ads for the Page.
  • Insights-See how the Page, content and ads perform.
  • This person can manage everything you can, including the ability to give access to others, remove anyone from the Page (including you) or delete the Page.

And some other actions that still seem to be in a bit of flux, you can now control a bit more than you could last year. Last year it appeared that anyone with full access had full control of everything and you couldn’t change it, now it appears you can set access levels with “Facebook Access” and they have added some additional options.

Change Access Level Image for Facebook

Task Access:

  • Community Activity- Review and respond to comments, remove unwanted comments and report activity.
  • Messages-Respond to direct messages as the Page.
  • Insights-See how the Page, content and ads perform.
  • Ads-Create, manage and delete ads for the Page.

Community Managers:

  • Community managers can moderate chat comments, suspend or remove people who violate community standards and see all admins of this Page.

 

As Facebook continues to make changes to the “New” Facebook experience, I would make a note to keep checking access levels and who has access to what. When the changes first rolled out last year, anyone who had any kind of access level to a page automatically got bumped up to “full” access. It appears (I hope) that they have fixed this and added additional levels and options, which is great.

 

What is not so great is that anytime Facebook makes a change, especially to options in the administrative section, they don’t tell anyone about it. I’ve been tracking the changes with the roll out to the new format since last year, and as usual, anytime a minor (but sometimes very important) change is made to business pages, they don’t make any kind of point of informing users about it.

 

Even worse is it appears they have now totally deleted the holding period for deleting a page (definitely moot if you don’t have access to it anyway, but….). In the old version, You’ll have 14 days to restore it in case you change your mind. After that, your Page will be permanently deleted. 

Old Deleted Page Image Facebook

Old Page Version

In the new version it looks like once it’s deleted, it’s gone. 

New Pages Experience Image Facebook

New Pages Experience Version

While it’s a lot harder to actually dig down to where to get to delete the Page in the new version vs. the older version, this is still NOT a good thing. So please stay on top of your access levels!

 

Tied into this, you think your personal Facebook account got hacked because you started getting reports from your friends that you were sending weird requests or odd messages to them.

 

There is a huge number of fake Facebook accounts, far more, I think, then Facebook will ever admit to. With the fake accounts, they take the name of someone, create a new account, and then take the profile picture and header image from the person they stole the name from and use it on the fake account. 

 

They then target your friends (because your friends list is open to anyone logged into Facebook) and start sending them friend requests. Many people accept the request because they see a name and photo of someone they recognized, so they don’t necessarily remember if they were already connected and hit accept. This is the way these fake accounts spread. Most people automatically assume they have been hacked, panic, and change their password. No, folks your account has not been hacked; it’s been cloned. Have friends report the fake profile ASAP. You need to lock down your friends list and also be very aware of what you post.

 

How can this hack your business account? It usually doesn’t directly, but it can cause identity theft of both your information and your friends’ information, and an awful lot of people have gotten scammed from these both identity-wise and financially.

 

And I do know two businesses that had employees that got their employee’s Facebook personal accounts cloned. The fake accounts messaged the owner of the business pages they worked for, and the business owners granted the fake accounts access to the business page, thinking they were the real employees and then had hacked pages and deleted pages. One more reason to limit access in the backend of Facebook.

 

Locking down your friends list and near the end (because it’s in the same section) how to see what apps have access to your personal (and business account with that).