In the ever-evolving world of technology, artificial intelligence (AI) has undeniably been the dominating force in 2023. This dominance is reflected in the “word of the year” lists compiled by renowned dictionaries. However, what’s noteworthy is that all the AI-related words highlighted on these lists are existing words that have been given new meanings, which some might find a touch predictable.
Key Takeaway
Artificial intelligence (AI) has made a significant impact on the “word of the year” lists compiled by Oxford, Cambridge, and Merriam-Webster. While the highlighted AI-related words are existing terms with new meanings, they reflect the growing influence and integration of AI in various aspects of society. By acknowledging these AI-related words, dictionaries emphasize the need to comprehend and adapt to the ever-evolving language of technology.
Cambridge’s Word of the Year: “Hallucinate”
Cambridge University has chosen the word “hallucinate” as its word of the year, which perfectly captures the habits of generative AI models like ChatGPT. These models have the uncanny ability to conjure up entire fictional personas and fabricated details in an attempt to avoid admitting any knowledge gaps. The irony lies in the fact that these systems are oblivious to their own ignorance.
Generative AI models prioritize producing sentences that resemble the patterns in their training data. So, if you ask such a system for information about famous 18th-century German surgeons and it has no direct matches, it will simply “hallucinate” something plausible but fictional, like Arman Verdigger from the Einschloss Research Hospital in Tulingen. It’s a skill that many of us could mimic, as long as it sounds convincing enough. Regrettably, these AI-generated hallucinations have been mistakenly accepted as genuine on numerous occasions.
That being said, hallucinations hold potential value. Generative AI models also produce images and audio that can be considered “hallucinated,” as they are essentially a mix of the model’s training data without being an exact recreation of any specific examples (although they can come startlingly close). This has led to the proliferation of AI-generated art and photos across various domains, with varying degrees of quality.
Henry Shevlin, an AI ethicist at Cambridge, believes the acceptance of the word “hallucinate” in its AI-related context reflects our readiness to attribute human-like qualities to the artificial intelligences we are creating. He predicts that as this decade progresses, our psychological vocabulary will further expand to encompass the peculiar abilities of these new intelligences.
Merriam-Webster’s Word of the Year: “Authentic”
Merriam-Webster, on the other hand, has selected “authentic” as its word of the year. While the dictionary hasn’t introduced a brand new definition for the word, it has assigned it a new and significant connotation. With the rise of artificial intelligence and its impact on deepfake videos, contractual issues in the entertainment industry, academic dishonesty, and a myriad of other topics, the line between what is “real” and what is “fake” has become increasingly blurred.
Authenticity has long been a concern, particularly in the context of modern consumerism. It is a paradoxical quality that cannot be bought or sold, making it perhaps the most valuable and marketable attribute in the world.
In the past, we worried about whether a trend or product truly represented the genuine interests and choices of individuals or groups. Now, we must grapple with the question of whether something is real in the first place, much like the Pope’s fabulous Balenciaga puffer – is it genuine or not?
The term “deepfake” also made Merriam-Webster’s longlist, transitioning from its unsavory origins in revenge porn to becoming a widely used term for generative AI. Although its roots may not be esteemed, we have little control over which words enter the zeitgeist.
Oxford’s Word of the Year: “Prompt”
Oxford University’s word of the year, unfortunately not AI-related, is “prompt,” which is a versatile and underutilized word. However, it has gained a new definition linked to the human side of generative AI.
When we instruct an AI system to generate a list of article ideas based on the current weather, for example, we are providing the “prompt.” This word has quickly transitioned from being primarily a noun to becoming a verb – now, we “prompt” the system.
These expanded definitions of “prompt” align quite naturally with its existing uses. After all, humans have been prompting responses for centuries. Interestingly, in the realm of computer interfaces, the use of “prompt” as a command line interface originated from the system prompting the human for a response. This reversal of roles prompts us to question who or what is truly driving the prompting. Whether this has enriched or diluted the word is subjective.
While “rizz” is not Oxford’s word of the year, it deserves a mention, serving as a playful abbreviation for “charisma” – something that AI arguably lacks, much like Tom Holland. It was only a matter of time before AI terminology infiltrated the lexicon, although it’s a bit disappointing that cooler terms like “latent space” have yet to enter mainstream usage. Nonetheless, as the technology continues to evolve rapidly, it may be prudent to stick with well-established terms for now, as determined by my esteemed colleagues in the lexicographic world. However, we eagerly await the future word choices of dictionaries, as bolder content teams consider giving a boost to terms like vectors and embeddings as well.
