Tired of the din caused by generative AI’s marketing hype? Though full of potential of AI, the technology has a long way to go.
The marketing hoopla around AI, and particularly generative AI (genAI), is growing tiring. You can’t read an article or watch a news broadcast without coming across at least one mention to it. We may be nearing the moment where we stop celebrating its advantages (while dreading some of its consequences).
The euphoria is so strong that a fallout, described by Gartner in its technology hype cycle assessments as the “trough of disillusionment,” is unavoidable and may occur this year. That demonstrates both genAI’s increasing promise and the technology’s immaturity.
The prospects for deep learning for predictive models and genAI for communication and content development are promising. But, notwithstanding the current marketing push, the difficulties remain stiff.
Machine learning techniques are only as effective as the data they are trained on. Companies are discovering that the millions of dollars they’ve spent on genAI have resulted in a poor ROI since their data contains inconsistencies, mistakes, and omissions. Furthermore, the technology’s excitement obscures the fact that many of the touted benefits are in the future rather than the present.
In summary, we’re not quite there yet, particularly with genAI-based chatbots, which have a tendency to “hallucinate” or crash repeatedly. Many genAI chatbots were recently revealed and are under fast development, despite being offered for beta-like general usage. And, honestly, the market is still figuring out how to effectively use the large language models (LLMs) that underpin many chatbots. (Learn more about LLMs below.)
Google, Microsoft, and OpenAI have hurried to build and distribute genAI technologies, but the result has been an unusual level of immaturity in many programmes. Chatbots generate content, but pinning your company’s reputation on the stuff they can provide right now might be career-limiting.
Here are some of the ways a genAI chatbot may get into trouble:
- Misinformation and misrepresentation.
- Deep Fakes (pictures, videos, voice clones), impersonation, synthetic pornography, and phishing
- Chatbot delusions and failures
- biases and inadvertent inaccuracies
- Concerns regarding copyright
- Potential government laws that might disrupt the genAI business.
- Investments that haven’t paid off as well as anticipated
- Overestimated productivity increases are not meeting business aspirations for outcomes.
Enterprises and business IT professionals should be experimenting with machine learning, deep learning, and genAI, but 2024 may not be the right moment for your firm to go all in. Wait for the election’s false news and potential misinformation to play out. Wait for the tools’ rough edges to be worked off, followed by more rounds of training. Wait for government rules to be implemented (or at least until you have a clearer understanding of what they want to regulate). If productivity is the aim, wait for others to experience the promised productivity benefits.
GenAI is still the next big thing, but it hasn’t evolved as much as the hoopla would have you believe.
Still confused about the differences between AI, generative AI, machine learning, and LLMs? The links below will help you get up to speed.
COURSES TO LEARN TO BE AI EXPERT
Machine learning
Machine learning is an area of artificial intelligence that employs data-driven algorithms to develop adaptable models capable of completing a variety of difficult tasks.
Deep learning
Deep learning is a subset of machine learning that use multiple layers of neural networks to tackle some of the most difficult ML problems without human intervention.
GENERATIVE AI
Generative AI, frequently referred to as genAI, is the artificial intelligence behind chatbots and other applications. It is a sort of artificial intelligence that creates pictures, text, videos, and other media in accordance with human directions.
Large language models (LLMs)
Large language models (LLMs) serve as the algorithmic underpinning for chatbots such as OpenAI’s ChatGPT and Google’s Gemini. An LLM is a computer algorithm that uses natural language inputs to anticipate the next phrase determined by what it has seen before. Then it guesses the following word, as well as so on until the result is complete.
