Lawmakers, tech experts sound alarm over artificial intelligence's potential dangers
Kara Frederick, tech director at the Heritage Foundation, discusses the need for regulations on artificial intelligence as lawmakers and tech titans discuss the potential risks.
Picture this: you open up your favorite food delivery app to order a late-night snack. You select your go-to order and finalize your purchase. When your food comes, you find that they gave you ranch dressing to go with your cinnamon roll. You know for sure, you asked for extra icing on the side and you check back on the app to find you indeed asked for icing, and received ranch.
You just experienced an artificial intelligence (AI) hallucination.
AI hallucinations are the category of content that may be inaccurate, nonsensical or even harmful due to AI models drawing information from outdated or incorrect data sets. In this case, the AI hallucinated what it thought would serve as a substitution for icing when the restaurant ran out. Without proper context, the AI did its best to work with what information it had.
(AI hallucinations are the category of content that may be inaccurate, nonsensical or even harmful due to AI models drawing information from outdated or incorrect data sets. (Getty Images))
The advent of artificial intelligence (AI) is unveiling numerous business opportunities. AI is predicting stock market trends, detecting fraud and malware before it can infiltrate systems and devices and communicating timely and helpful updates with customers. However, with new technology comes new risks that can be much more serious than mistaking ranch dressing for icing.
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
AI hallucinations are a new business threat that has risen to prominence due to the introduction of generative AI (GenAI). Skewed predictions and inaccurate data produced by AI hallucinating can put businesses’ and individuals’ reputations at risk due to poor decision-making or even copyright and legal issues as a result of the AI being trained on already existing data or content available in the public domain. Businesses need to ensure their AI technology is built on reliable models with access to fresh, continually updated data in order to significantly reduce hallucinations.
Generative AI in the Wild
Businesses in every industry are evaluating GenAI opportunities right now. Generative AI – AI technology that can generate various types of content (e.g. images, videos, audio, text and more) by leveraging prompts and existing data – can be used for industrial monitoring, medical devices, health care diagnostics and other endless possibilities.
It’s no surprise that IDC forecasts spending on GenAI solutions will hit $143 billion in 2027. A study conducted by Salesforce even found that 45% of the U.S. population is already using GenAI.
McDonald’s is considering automated voice ordering for drive-throughs, Stitch Fix is experimenting with GenAI to suggest styles for its customers, and Morgan Stanley is building an AI assistant with GenAI. People and organizations globally are accessing services every single day that use AI without even knowing it, which is why it is so crucial for businesses to build AI into their platforms in a safe, strategic manner.
Many use cases for AI require paramount or even life-altering decisions to be made in an instant, such as a medical diagnosis or decision in the middle of a surgery. Therefore, the accuracy and quality of data GenAI models are trained on are of utmost importance.
RESEARCHERS CAN'T SAY IF THEY CAN FULLY REMOVE AI HALLUCINATIONS: ‘INHERENT’ PART OF ‘MISMATCH’ USE
Data Has Expiration Dates Too; Real-time Data for Really Great Results
Food and gift cards expire, and data can too. What good is data if it's outdated and inaccurate? For GenAI to do its job – generating new data – the model it is built on needs current, contextually relevant information to learn from. Deep learning is a complex computing process used by GenAI models to analyze patterns in large data sets to create new outputs.
Data is the fuel behind AI, which is only as good as the data it’s trained on. The quality of data is increasingly important as sometimes, AI models may mislabel or wrongly categorize data, resulting in AI hallucinations. Businesses can mitigate this by incorporating real-time data.
CLICK HERE FOR MORE FOX NEWS OPINION
Real-time data is delivered immediately after it’s collected, providing a steady stream of information without delay. It can ensure that an AI model’s predictions are in sync with the most recent data produced. This type of fast-paced, always up-to-date technology can significantly reduce the risk of hallucinations and is therefore essential for businesses to properly leverage the full potential of GenAI to drive decision-making and produce positive business results.
Getting Ahead of AI Hallucinations Drives Positive Business Results
While businesses may not be able to eliminate AI hallucinations completely, there are necessary steps organizations should take to prevent them from happening to avoid costly risk and potential harm. GenAI has already shown us a glimpse of just how many ways our lives can change. When built on an AI model that leverages real-time data, AI can and will continue to be a part of our lives, providing improved services, faster response times and new ways to leverage technology.