How to discourse on Artificial Intelligence (even if you possess limited knowledge of the subject)
AI has become a hot topic of conversation lately. However, if you find yourself feeling overwhelmed or uncertain about the concepts being discussed, there’s no need to fret. I’m here to assist you.
To help you speak about AI confidently, I consulted some of the finest AI journalists in the field. My colleagues and I dedicate our days to immersing ourselves in the technology, attentively listening to AI experts, and then translating their language into clear, relatable terms with significant context. I dare say we possess a considerable amount of expertise in this subject matter.
Below, you will find seven key factors to consider when discussing AI.
7. Embrace the freedom to be uninformed
The tech industry often struggles to explain itself clearly, despite its insistence that large language models will revolutionize the world. If you find yourself grappling with the concepts, rest assured that you’re not alone, says Nitasha Tiku, the tech culture reporter at The Washington Post. Additionally, the presence of jargon further complicates conversations about AI. Tiku explains that hallucination is an elaborate way of saying that an AI system fabricates information, while prompt engineers are individuals who possess the ability to effectively communicate with AI systems to achieve desired outcomes.
Tiku suggests watching explanatory videos on platforms like YouTube to grasp various AI concepts and models. Opt for hosts who adopt a more subdued tone, such as those on Computerphile, instead of following AI influencers. If you prefer concise and straightforward content, IBM Technology is an excellent choice. While there isn’t a channel exclusively catering to casual observers, it can assist in demystifying the process.
Furthermore, it’s important to remember that regardless of how you discuss AI, there will always be some detractors. The world of AI seems to have fragmented into factions, with individuals talking past one another while clinging to personal definitions and beliefs, remarks Will Douglas Heaven, the senior editor for AI at MIT Technology Review. He advises you, to determine what AI means to you and remain steadfast in your interpretation.
6. Be precise in delineating the type of AI you refer to
‘AI’ is often treated as a monolithic entity in public discourse, but in reality, AI encompasses a myriad of disparate elements, asserts Karen Hao, the China tech and society reporter at The Wall Street Journal (and the creator of The Algorithm!).
Hao recommends distinguishing the specific AI function you are discussing to engage in a more nuanced conversation. Are you referring to natural language processing and language models, or perhaps computer vision? Alternatively, are you exploring different applications, such as chatbots or cancer detection? If you are unsure, Hao provides comprehensive definitions of various practical applications of artificial intelligence.
Billy Perrigo, a staff reporter at Time, concurs that talking about AI as a single entity obscures the true nature of the technology. He states Multiple models possess distinct capabilities, yielding varied responses to identical prompts, while each model also carries its own biases.
5. Maintain authenticity
The most crucial inquiries one should pose regarding new AI products and tools are simply: What can it do, and how does it achieve these outcomes? advises James Vincent, senior editor at The Verge.
There is a prevailing trend within the AI community to dwell on the long-term risks and potential of AI. It is tempting to be lured into hypothetical scenarios and envision what the technology might accomplish in the future. However, Vincent contends that discussions about AI are more fruitful when grounded in pragmatism, focusing on actuality rather than indulging in conjecture.
The tech sector has a tendency to overstate the capabilities of its products. Approach AI discourse with skepticism and cynicism asserts Douglas Heaven.
This approach becomes especially pertinent when addressing AGI, or artificial general intelligence, which typically refers to software possessing human-level intelligence (although what precisely constitutes human-level intelligence remains open to interpretation).
If something sounds like poorly conceived science fiction, it probably is, he adds.
4. Manage your expectations
Language models that power AI chatbots, such as ChatGPT, occasionally hallucinate or fabricate information. Although this can be disconcerting and surprising to users, it is an inherent aspect of their functioning, explains Madhumita Murgia, artificial intelligence editor at the Financial Times.
It is essential to remember that language models differ from search engines designed to provide definitive answers. Language models are predictive systems that generate the most probable words based on your query and the entirety of their training data, clarifies Murgia.
This does not imply that they are incapable of producing original content… However, we should always anticipate inaccuracies and the fabrication of facts. By doing so, errors become less consequential since our usage and their applications can be adjusted accordingly, she remarks.
3. Avoid anthropomorphism
AI chatbots have captivated the public’s imagination due to their ability to generate text that closely resembles human writing. They create an illusion of interacting with something beyond a mere computer program. However, it is crucial to refrain from attributing human characteristics to this technology, warns Chloe Xiang, a reporter at Motherboard. She advises against using gendered pronouns or ascribing feelings, thoughts, or beliefs to AI.
Engaging in anthropomorphism perpetuates the misconception that AI systems possess greater capabilities and sentience than they actually do.
Personally, I often find myself prone to this error since our language has yet to catch up with terms that accurately describe the actions of AI systems. When in doubt, I substitute AI with a computer program. Suddenly, it feels rather silly to claim that a computer program advised someone to divorce their spouse!
2. Recognize the power dynamics at play
While the hype and nightmarish scenarios often dominate news headlines, it is crucial to consider the role of power when discussing AI, asserts Khari Johnson, a senior staff writer at Wired.
Power is a fundamental element in the ingredients required for creating AI, including computational resources and data. It is also vital when questioning the ethical usage of AI and understanding who can afford to pursue advanced degrees in computer science, as well as who participates in the AI model design process, Johnson emphasizes.
Hao agrees, adding that it is helpful to bear in mind that AI development is profoundly political, involving massive financial investments and numerous factions of researchers with conflicting interests. She notes, Sometimes, the discourse surrounding AI is less about the technology itself and more about the individuals involved.
1. Please, for heaven’s sake, let go of the notion of robots
Do not envision or describe AI as a terrifying robot or an all-knowing machine. Remember that AI essentially boils down to human-coded computer programming, achieved by combining extensive datasets with considerable computational power and intelligent algorithms, advises Sharon Goldman, a senior writer at VentureBeat.
Read More: Introducing the Redmi Note 12T Pro: Unveiling Specifications, Pricing, and Availability
