Perils of AI for small business

  • Post author:
  • Post category:AI

Perils of AI for small business: you need to protect your business from the AI gold rush. You can mitigate these perils with rules for employees and purchasing.

AI Gold Rush

There’s gold in them AI hills. Seems like every huckster, hypester and fraudster who was promoting BitCoin and VR have rushed into the AI gold rush. Recently the Federal Trade Commission in the U.S. warned it may crack down on companies that not only use generative AI tools to scam folks, but also those making the software in the first place, even if those applications were not created with that fraud in mind. An example would be using a chatbot to produce a malicious computer virus.

Microsoft was in such a rush to cash in they disbanded a key AI ethics team before product launch.

Here are some of the main perils for small businesses with chatbots and AI.

Chatbots guess the answers

Chatbots (aka Long Language Models or LLMs) guess the answers to questions. It’s fake it ’til they make it. Also the data you feed the AI can limit the results. See Here’s how the data we feed AI determines the results.

CNET stopped using AI generated content back in January due to the mistakes in AI articles about financial planning.

AI generated code still needs to be checked for security and other errors because it’s a guessing game.

My favourite example of what can go wrong is “We asked ChatGPT to write our Christmas Cards. It didn’t go well.” Besides the pictures being wonky the messages inside the cards were inappropriate, banal or creepy.AI Xmas card

Sensitive Data Loss

Less than three weeks after Samsung lifted the ban on employee use of ChatGPT, staff at Samsung subsequently dumped into ChatGPT corporate secrets, including equipment measurement and yield data from the conglomerate’s device solution and semiconductor business unit.

OpenAI only recently put in new controls for ChatGPT that allow users opt out of providing conversation history as data for training AI models.

AI Treasury of Hackery

AI also provides a treasury of hackery. First, it is now very easy for people with no programming skills to produce malicious viruses. Second, hackers can use AI to generate better phishing emails. Third AI can be used to produce really good looking fraudulent résumés, cover letters, reports, graphics that can fool a business owner.

Currently there is no software that can detect AI fakes, particularly written fakes, reliably.

Mitigate the perils

How do you mitigate these perils?

First, set up clear rules for employee use of chatbots and other AI tools in the business. Focus on keeping business secrets away from chatbots and accuracy of the results. For another example of accuracy problems with AI, read The World’s First Robot Lawyer Isn’t A Lawyer, And I’m Not Sure It’s Even A Robot.

Second, be wary of the gold rush sales pitches around AI. Tech companies are rushing out chatbots or rebranding existing products as AI driven to drive up sales with little concern for security, accuracy or even longevity. I mention longevity as a lot of these free or low cost AI products are sure to disappear after a year. AI requires very expensive and vast computing resources. No one but companies like Microsoft or Google have the computing and financial resources to offer AI at low prices.

Third, I expect to soon see new federal/provincial laws governing AI and also professional society rules. Already a judge has ruled the even under the current law you cannot be represented by a robot or AI lawyer in court. Watch for those rules and how they affect your business.