This is the eighth article in a series about artificial intelligence (AI) written over the last couple of years. While the previous features have attempted to educate on what AI really is, and what different versions of the technology exist, the aim of this article is to debunk common misconceptions and myths about AI.
Misconceptions Plaguing AI
The peak misconception is that AI “thinks” or behaves like a human brain, possessing true understanding, consciousness or intent. In reality, AI systems are no more than sophisticated statistical algorithms that recognize patterns and calculate probabilities to generate responses.
At this stage, AI is not truly comprehending, experiencing or innovating like humans.
The misconceptions often arise from anthropomorphizing AI — attributing human characteristics to machines and often lead to unrealistic expectations as well as unfounded fears.
As one expert opines, “Misconceptions are due primarily to media hype, and limited understanding of the technology’s actual capabilities.”
Key misunderstood areas of AI include:
- AI does not have human-like reasoning. Instead, Large Language Models (LLMs) predict the next likely word in a sequence based on statistics, without understanding meaning or intent.
- AI lacks emotions, self-awareness and motivation.
- AI is not necessarily objective. Its behavior depends on the data it was trained on and functions prescribed by creators.
- AI is not perfect. It can produce “hallucinations” defined as confident but false information derived from their training data.
- AI requires clean, relevant data to be effective, and more data does not always equate to better results. If the AI training data is flawed or skewed, the output will be as well.
- AI is not one thing. It ranges from simple chatbots to advanced machine learning systems.
Unfounded Fears of AI
- Fear: AI will replace all human jobs
- Fact: AI automates repetitive tasks and is designed to augment human intelligence.
- Fear: AI operates autonomously without human oversight
- Fact: AI systems require human intervention to validate outputs, especially in complex or high-stakes scenarios (known as “human-in-the-loop”).
- Fear: AI is “intelligent” and can solve any problem
- Fact: AI is highly specialized, or “narrow,” and is designed to perform specific tasks.
- Fear: AI “learns” on its own
- Fact: AI requires humans to set up training datasets, define problems and continuously update software.
- Fear: AI is out of control
- Fact: AI is not taking over, but it is changing how we compete and grow at a rapid rate.
Will AI Become Sentient?
Perhaps the biggest debate and fear for many is that AI will become sentient.
However, what we do know is that AI is currently far from achieving artificial general intelligence. It doesn’t have the creative abilities or emotional intelligence of people. Most AI models are designed for specific tasks and are constrained and limited by those parameters.
The consensus of subject matter experts is that “The development of artificial intelligence is a rapidly evolving field. And while progress is inevitable, we are unlikely to see sentient AI any time soon.”
Closing Thoughts
At the core, AI systems are only as good as the goal of the creators and the data they train on. Moreover, AI will not make human labor obsolete. As we have experienced over many years, new technologies have increased productivity, created new jobs and new industries. This is also true with AI.
As AI adoption continues, it will require new skills to perform newly defined tasks. This will undoubtedly cause many jobs to shift, especially those in the business world that are repetitive in nature. However, workers whose jobs are shifted or displaced will require new skills.
In summation, the call to action is to understand AI, identify where it fits in your business model and then equip people to allow for stable careers in a shifting landscape.
Alan C. Brawn, CTS, DSCE, DSDE, DSNE, DCME, DSSP, ISF-C, is principal of Brawn Consulting.
