Latest News

.: AI Updates & Developments

​

.: Google Announces ‘Assistant with Bard’ for Android and iOS ➜ A combination of the generative and reasoning capabilities of Bard with the personalised help of Google Assistant. This includes Bard Extensions that can access Gmail, Google Drive, and Docs to answer queries. Additionally, Assistant with Bard has a “conversational overlay” that can accept text, voice, or image input. Google calls this an “early experiment,” with plans to roll it out to early testers for feedback before public availability over the next few months.

.: AI Facial Recognition Wrongfully Imprisons Innocent Man ➜ In a landmark incident, Robert Williams was wrongfully arrested in January 2020, marking the first documented case in the U.S. where facial recognition technology led to a false detention. This arrest occurred amidst a surge in law enforcement’s use of powerful AI for facial recognition. Williams’s case, resulting from a mistaken match by the Detroit Police Department’s facial recognition system, underscores the emerging challenges and ethical considerations in deploying AI technologies within the criminal justice system​. Despite the known flaws and the potential for mass surveillance threatening privacy, law enforcement continues to increasingly rely on such AI systems.

.: ‘Alarming’: Convincing AI Vaccine and Vaping Disinformation Generated by Australian Researchers ➜ Australian researchers have highlighted the power of AI to generate harmful disinformation. In an experiment, they used AI to create over 100 misleading health blogposts in multiple languages within just over an hour, bypassing safeguards meant to prevent the generation of misleading or harmful content. The experiment underscores the need for stronger industry accountability and better safeguards against the misuse of AI.

.: Why are Fewer Women Using AI than Men? – BBC News ➜ The article explores the reasons behind fewer women than men using artificial intelligence (AI), particularly AI chatbot ChatGPT. While the chatbot has over 180 million users, many women, including jeweller Harriet Kelsall and business coach Michelle Leivars, express concerns about the reliability of the AI and the potential loss of authenticity in their communication. A survey earlier this year revealed that only 35% of women use AI in their professional or personal lives, compared to 54% of men. The article suggests that this disparity may largely be due to the confidence gap and the fear of criticism that many women face when using AI tools.

.: Most of our friends use AI in schoolwork – BBC News ➜ A recent report by BBC Young Reporters Theo and Ben explores the use of Artificial Intelligence (AI) among pupils in their school. The majority of their peers admit to using AI, specifically ChatGPT, to assist with homework, formulating ideas, and structuring their work. However, some students confess to the misuse of AI in providing answers, a practice that has resulted in inaccurate information. Despite these drawbacks, many still find the AI tool useful and suggest it should be taught in schools.

.: The Screen Actors Guild’s strike-ending deal has entered its final step ➜ The Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA) has reached a tentative deal with Hollywood studios, ending a 118-day actors’ strike. The agreement, approved by the national board, is awaiting final ratification from union members. A significant aspect of the deal is a set of protections around the use of artificial intelligence, which mandates informed consent and compensation when guild members are replicated digitally using AI.

Reflection

.: Why this news matters for education

​

Although the Humane AI Pin is a curiosity, when you look a little more closely, it is just a phone without a screen. It even comes with a monthly plan from T-Mobile in the US!

Also, the demo video included a glaring error about the best place in the world to see the equinox. Another example of hallucinating large language model results added to tech demos without fact-checking. I am looking at you, Google Bard! (through a telescope 😉 )

We already have very powerful devices in our pockets managed differently by schools and education systems. The mobile infrastructure is vast, on which we might experience AI-augmented learning.

According to a report from Statista, there are more mobile subscriptions than people on the planet.

So, I wonder how we leverage our devices in better ways to assist, augment and amplify teaching and learning.

The path ahead of personal devices, whether smartphones, pins or glasses, making it easy to access advanced AI capabilities matches another direction. The personalisation of AI implementation through agents, designed for narrower tasks – powered by a richer context of who you are.

All of this is glued together with data, making me wonder: who owns my heart rate data from my Garmin watch?

Which activates the part of the data ecosystem around fitness and health information gathered, stored and analysed by various apps and wearables. How might we connect these data pools to further enrich the learning experience?

.:

​

~ Tom

Prompts

.: Refine your promptcraft

​

This week I am sharing my draft of an Imaginary Scenario Prompt Framework. The aim of this multi-step prompt interaction is to surface assumptions and constraints and then scaffold thinking that pushes beyond those limitations.

A pre-requisite is a chat session where you have been building, designing and exploring some new ideas. Use this set of prompts once you have a conversation to review.

1. Constraint Analysis Prompt:

Aim: Concisely identify the core constraints for the LLM to focus on.

2. Imagined Future Prompt:

Aim: Spur your own creative thinking about an optimistic but grounded future where key constraints are lifted.

3. Follow-up Prompt:

Aim: Iterate for more details and depth on the imagined scenario.

.:

Just a post script that statistical language models have limits to how creative they are. (If they are creative at all!) They are built to predict the most likely word, rather than diverge to something unexpected, so keep that in mind.

My approach is to collaborate with a wide range of AI tools to amplify my creativity, not to sit back and think an LLM can do better. I encourage you to stay in the creative loop.

And, as I always say, remember to mak​e this your own, tinker and evaluate the completions.

Learning

.: Boost your AI Literacy

​

ARTICLE .: Untangling AI Hype from Reality ➜ This ABC News article demystifies the hype surrounding AI by exploring its current capabilities, limitations, and future potential. It offers a grounded perspective on the state of AI technology, making it an essential read for those looking to understand the realistic prospects of AI.

COURSE .: Unlock AI Secrets with Amazon’s Free Learning Resources ➜ Amazon’s initiative, as highlighted by ZDNet, opens doors to free AI learning resources. It’s an excellent chance for educators and learners to enhance their AI skills and knowledge without the financial barrier, fostering broader accessibility to AI education.

EXPLANATION .: Explained: Generative AI ➜ MIT News provides an insightful and accessible explanation of Generative AI, a crucial domain within the AI landscape. This resource breaks down the concept, its applications, and significance, making it a valuable educational tool for anyone interested in this aspect of AI.

Ethics

.: Provocations for Balance

​

• With the rise of emotional analysis AI, how do we protect people’s psychological privacy? Should individuals have a right to consent before their emotions are analysed by algorithms?
• If an AI system makes a mistake that harms a student’s learning or future prospects, who is liable? How do we balance accountability with encouraging innovation in AI for education?
• Should educators be required to disclose when AI is being used for certain teaching tasks? What happens when it swings the other way, and we have less trust of human-only generated content?

~ Inspired by this week’s developments.