There Isn’t Much I In AI
Silicon Valley has a serious problem with fraud right now. Whether it’s tech companies lying about what they’re actually doing to their investors, blockchain tech lacking a lot of basic security features while touting greater security, that whole NFT nonsense, or rebranding already existing inventions to pretend they are great innovations, there’s a lot of nonsense happening in the tech industry. AI is the newest thing they’re trying to push.
Much like the blockchain, AI is a solution in search of a problem. Companies are sinking billions of dollars into developing it, promising it can do anything, but companies that implement it are finding it’s not as useful as they’ve been told. When it comes to businesses, especially local businesses, AI is not your friend. Here’s why.
The Problems With AI
1. It’s Not Intelligent
Despite Google employees declaring that their programs have obtained sentience, it’s fairly easy to see that they haven’t. For one, Google’s attempts to implement it into their search functions has been disastrous. Just give this a watch!
What I sometimes think even the tech industry doesn’t understand is that computers don’t actually think, they just follow instructions. Through large language models and algorithmic coding, you can create very complex instructions, but they’re still just instructions. The machine isn’t thinking, it’s just scraping data it’s been fed and then determining what to output based on those instructions. That’s why you get Google’s AI is telling you to eat rocks and smoke cigarettes while pregnant. It doesn’t know how to think about the information, it can only tell that the information matches enough of the question to be relevant.
2. It Can Make Stuff Up
Techbros have promised that AI can write articles for you, but there’s a problem. If you want to avoid copyright lawsuits, you can’t just have the AI regurgitate sections of text it’s crawled word for word. It has to alter the text enough to not match the source. Doing that sort of rephrasing requires a grasp of context and creativity that is simply beyond AI. When you program a large language model to be able to alter the text it scrapes to create new text, you get a problem that they have dubbed “hallucinations.” The algorythm will mash text together in a way that makes syntactic sense (it’s grammatically correct), but it has problems connecting ideas and will sometimes appear to just make things up.
For example, the YouTube user Whirly made an AI generated documentary on the megalodon. Some highlights include, claiming it was a land animal, that it was more closely related to the jaguar than to actual sharks, that it primarily ate humans, and that someone named David Wellington is somehow involved.
If you use these AI to generate materials, or to act as a chatbot for your website, the business owner is liable for anything it says to customers. Do you want your business on the hook for a program that will sometimes make things up out of whole cloth? I don’t think so.
3. It’s Shockingly Easy To Bypass Safety Parameters
Of course Google knows it’s a bad idea to sell an AI that might make up stuff, especially if that stuff is libelous or offensive. To fix this, they’ve put in safety parameters to stop it from doing so. Problem is, becauseit doesn’t have the ability to think, it’s very easy to get around them. The AI can be programmed not to lie, but if you tell it to create a hypothetical or fictional scenario, it will completely ignore all of those safety parameters and do it anyway. All Google can do to fix this is try to code in more parameters, but the end result of that will be that it obstructs that program’s ability to generate content, defeating the stated purpose of the program.
4. The Well Is Poisoned
People have so many different reasons for hating AI. Some of those reasons are nonsense, such as believing it will gain sentience and become a threat to humanity. Some of those reasons are personal, such as artists fearing businesses will fire them because the AI is cheaper. Some hate it simply because it has been used in a lot of fraud. Whatever the reason for it, however, there are a LOT of people who hate AI. Now that there have been so many widespread attempts to use it, people are starting to pick up on what AI generated content is like and hate it on principle, and hate any business that uses it. Regardless of Silicon Valley’s stated potential for the technology, it’s been implemented so poorly that the negative impression has poisoned the well, as the saying goes. People don’t want anything to do with it, and if you use it, they won’t want anything to do with you.
While not everyone is ardently against this technology, there’s a large enough population of people who will reject it that it can be a serious risk to your business to employ the technology. For the time being – and probably the foreseeable future – it’s probably in your best interest to avoid it. If you’re running a local business, your market is probably too small to run the risk of the technology turning anyone away.
Photo by Matheus Bertelli
5. It Sells You Solutions To Problems It Creates
AI has been promoted as a solution to so many issues, but in some cases it creates those solutions in the first place. This video here is the perspective of one coder who watched the process of hiring people be utterly destroyed by AI. First, people would use AI to write resumes and cover letters that would be sure to hit all the keywords that businesses would be looking for. The ease of this resulted in businesses getting flooded with so many resumes they couldn’t go through them all. So the same company sold their software to businesses, promising that it would help them filter through resumes to find good ones. The problem was that the algorithm was just searching for the same specific keywords and phrases as it was writing the resumes for in the first place. The end result was that the only resumes that got filtered out were the ones that people actually put thought and effort into, leading to only lazy people getting past the first cut. Not the best pool to be drawing from.
6. It Disrupts It’s Own Business Model
I mentioned early that Google built AI into its search engine with less than stelar results. Well, even if you discount the nonsense answers it spits out, this creates another problem. Google is becoming less and less reliable in the eyes of its users. What made Google so popular in the first place was how it made sure to verify the results it fed you so that it served you up the most relevant and most accurate results before anything else. Getting on the first page of Google search results was a sure-fire way to have your business grow its customer base.
Now, getting on the first page of Google is next to impossible. The AI search function and the several ads that it shoves into your face eat up all the first page real estate. Even if you are the first legitimate result in the search engine’s algorithm, you might still get shoved onto page 2 of the results to make room for everything else.
On top of that, if the AI results were ever corrected to avoid all the problems above, it would still drive traffic away from your business. The AI results would scrape text off your website, providing the answers that the searcher is looking for. Since they already have the answer, why should they bother going to your site in the first place? It steals clicks from you, completely undoing all your efforts to be found online. Because of all this, it seems inevitable that Google’s dominance over the search engine industry will break unless they reverse course.
Don’t hold your breath waiting, though.
What Can You Do Against AI?
That’s a pretty good question. At least for the moment, Silicon Valley seems insistent on trying to shove this down our throats. What can you do? Atlas Internet Marketing has a few suggestions.
1. Try UDM14
The sheer loathing for Google’s AI results has prompted people to find a way around them. One clever person found out that you can evade them by using Google’s testing code that turns the AI results off. The code is &udm=14. They put this into a website that appends this code to the results and then serves up the Google results with all that nonsense cleaned out. Check it out: https://udm14.com/. Of course, this only works for your own searching, but you can help spread word of it so more people use it.
This isn’t a very effective method, however, so there are some other things you should do.
2. Don’t Be Entirely Reliant On SEO
Some marketing companies will sell cheap marketing packages for only a couple hundred dollars a month. This sounds like a saving, but these packages only include basic SEO. AI is beginning to render that less meaningful. While it still has some uses and should be implemented into your website, you need a more layered approach. In fact, Atlas says that you really should never have been completely reliant on it. A good marketing strategy will incorporate a variety of approaches. The following examples are some of the most effective.
3. Focus On Reviews
Word of mouth is still the best marketing you can get. It remains absolutely true and one way the internet has made this a more viable strategy is reviews. People will look for online reviews when deciding what companies to go with. Asking for reviews is easily the most effective way to get them. If you provide a good customer experience, people are more likely to give you a good review when you ask for it. These can be reviews on your website itself, or through facebook, or other review platforms. The more positive reviews that can be found, the more likely people will be to give you a try.
4. Follow Up With Customers
It’s easier to keep old customers than find new ones. This should give you an incentive to keep them. For small businesses, it’s not that difficult to follow-up with customers. With text and e-mail, you can have them opt-in to communications. Follow-up after purchases to make sure they’re satisfied with the results (and give yourself an opportunity to remind them that reviews always help). Send them newsletters with reminders of new products or services, discounts, and coupons. Loyalty programs for regulars will encourage them to come back again and again. This helps build a strong relationship with the people who keep you in business.
Be careful not to overdo it, though. A huge newsletter every month is more likely to annoy customers than encourage them. Keep the newsletters brief, with a few important updates and offers, then point them back to your website for more details.
5. Ask For Referrals
Your loyal customers are the ones who will provide you with the word of mouth you need for this strategy. Once you’ve got that loyalty, you want them to share their experiences. Reviews are an important part of that, but referrals are also really good. If you can get a customer to convince even just a few friends to try your business, that’s a significant opportunity to grow your customer base. Just like with reviews, all you need to do is ask.
As with everything, be sure to thank your customers for their support. Not just with newsletters and emails either. Customers know that a lot of these will be general and/or automated, which can make it less meaningful (though still effective). Make sure to thank every customer in person when they do business with you, especially the repeat customers. Here, loyalty programs will help you even more, as having customers sign up for them will not only let you know who’s coming back regularly, but also help you familiarize yourself with their names and purchase history. You can tailor your interactions with them to be more specific, which makes them feel really appreciated.
Real People Get Real Customers
Photo by Pavel Danilyuk
AI has a lot of problems right now. The common response to this is “it will get better,” but how dependable is that? “It will get better” has been the refrain of so many Silicon Valley tech scams; a catch-all defense against any criticism. Even just a cursory glance through the list of tech industry frauds in the last ten years should tell you there’s no promise there. That said, even if AI could live up to its own hype, there’s nothing better than a real human encounter for building customer relationships. Machines can automate a lot of mundane and routine processes, but talking to someone in person will always be more satisfying. While it can be hard to convince customers that a business cares about them, they know for a fact that a machine doesn’t.
Success in business is building up a loyal customer base and AI can’t do that. There may be some use cases for the technology, but they will be niche and like any tool will require you to train your employees (and yourself) in their quirks in order to effectively use them. They will never be a true replacement for people. You will need to think carefully about what the AI can actually do (which will require some good research first) before you can decide how, or even if, you want to implement it in your business.
While it may have some uses, at present the ways that it can harm your business are much more prevalent. Because of that, it’s fairly safe to say that, right now, AI is not a friend to your business.
