Every day, a new headline emerges about artificial intelligence (AI), promising either salvation or doom. From “AI will replace all jobs” to “AI will cure cancer,” the relentless marketing of AI solutions can make it seem like we’re on the brink of a utopia—or dystopia—depending on who you ask. But amidst all the buzz and excitement, a lot of it is, frankly, bollocks.
Continue reading
This article explores the impact of AI on creative professions, using comic book artists as an example. It discusses concerns over ownership, consent, and job security, while also highlighting AI’s potential to enhance creativity and streamline workflows. By comparing these effects with those in copywriting and programming, the article provides insights into navigating the challenges and opportunities AI brings to the creative industries.
Continue reading
This article explores cognitive dissonance in AI, focusing on the inconsistencies in AI system outputs due to conflicting data, rules, or patterns. It examines key issues such as drift, bias, overfitting, underfitting, explainability, transparency, data privacy, security, hallucinations, and data poisoning. The article also provides strategies for addressing these challenges, emphasizing the importance of continuous monitoring, bias mitigation, model complexity balance, enhancing explainability, robust data governance, and protection against data poisoning. The goal is to develop more reliable, fair, and trustworthy AI systems.
Continue reading
Artificial Intelligence (AI) language models often generate different responses to the same query, leading to perceptions of inconsistency and subjectivity. This article delves into the reasons behind this variability, including the probabilistic nature of AI, contextual dependence, diversity in training data, and other influencing factors. It also offers insights on achieving greater consistency in AI interactions.
Continue reading
FICTA 24 brought together experts to discuss AI’s transformative potential and the challenges ahead. My keynote, “AI: the spearhead of progress, not quite there”, covered the current state of AI, its applications across various industries, and the ethical considerations crucial for its future. The conference underscored the importance of interdisciplinary collaboration and featured esteemed speakers who provided valuable insights into AI and technology.
Continue reading
As part of the Cheltenham Literature Festival 2023, in honour of Ada Lovelace Day and as a reaction to a new report from the University of Bristol’s Research Institute for Sociotechnical Cyber Security (RISCS), The Times hosted a discussion panel “What If AI Doesn’t Change the World”. Led by The Times technology business editor Katie Prescott, and with Cambridge Professor of Politics David Runciman, Oxford AI systems expert Michael Wooldridge and AI ethicist Kanta Dihal, looking to explore the promise and peril of AI, asking whether our fears for the future are in fact misplaced? The audience was invited to ask the panel questions, but invariably, they didn’t have time to answer them all. As well as document the event, I thought it would be fun to ask ChatGPT what they thought…
Continue reading