AI can read your thoughts!

A story published in the Guardian is fascinating. Researchers at The University of Texas, Austin using LLMs (like ChatGPT), have been able to reconstruct stories from fMRI scans of peoples' brains. This has enormous potential benefits for people suffering from strokes and other neural-cognitive issues. But ethically, it's a potential minefield. Imagine if China could read the thoughts of its people? 

Blog reawakening

 This blog is coming back to life. I shut it down at the end of 2020 when I retired for health reasons. I report that my health is much better (though not perfect) now. However, since 2022 with the advent of ChatGPT and AI being in the news constantly, it seems timely to revive this blog. I've been working in AI since starting my MSc degree in 1985, so to leave the field just as it finally becomes significant seems crazy. This blog will therefore concentrate on the history of computing, focusing now on AI and its potential impact. That is what my book The Universal Machine does after all.

Blog Closing Down

Hi,

I'm closing this blog and this will be the last post. I started it with the first post on April 5th 2010 to support the writing of my book, The Universal Machine, on the history of computing. Over the years it's by and large kept to that topic. It's had 585,622 page views over the decade or so which sounds quite respectable. The blog will remain in place as a record, but I will not be adding new posts as I'm retiring at New Years. Thanks for your support.

Ian Watson

Enigma machine found in Baltic to be restored

 

Divers who found a WWII Enigma coding machine in the Baltic will have their find restored in a German museum, the process will take about a year. The enigma machine is a German Naval Enigma with three coding rotors, not the more complex four-rotor machine used by the U-boat service. The Enigma machines were used throughout WWII but had been cracked by Alan Turing's team working at Bletchley Park in England. As a consequence, the British could read all the German navy's communications from 1941.

Final 2020 Gibbons lecture - 21st October

 The fourth and final 2020 Gibbons Lecture is on Wednesday 21st October 2020 from 6-7pm in Room B10, The General Library, 5 Alfred St. Refreshments and networking from 7pm.  

Please register here to attend in person, if you have not already done so.   This lecture will not be live-streamed nor available later as a podcast.

 

Social media, AI, and society: Some psychological insights

As advanced digital technologies become an indispensable part of nearly all aspects of everyday living, it is essential to consider the downstream effects on society.

In this talk, Associate Professor

Kumar Yogeeswaran from the College of Science at the University of Canterbury will consider social scientific evidence on how these indispensable technologies shape people’s attitudes, beliefs, and behaviours. 

This lecture is run in association with the Auckland ICT Graduate School, IT Professionals NZ, and Dr Beryl Plimmer.

See here for more details on the Gibbons Memorial Lecture Series.  

Third 2020 Gibbons lecture - 14th October

The third 2020 Gibbons lecture is on the 14th October 2020 from 6-7pm in Room B10, The General Library, 5 Alfred St. Refreshments and networking are after the lecture at 7pm.  

Much of scientific research involves the generation and testing of hypotheses that can facilitate the development of accurate models for a system. In machine learning the automated building of accurate models is desired, however, traditional machine learning often assumes that the underlying systems are static and unchanging over time. In this talk, Dr Yun Sing Koh from the School of Computer Science at The University of Auckland will discuss research in the area of data streams and how we adapt to changes in the data.

Please register here to attend in person. The lecture will be streamed live and later available as a podcast.  

Gibbons Lecture Series - Big data: Transparency and reliability

The second 2020 Gibbons Lecture will be live-streamed only on 7 October 2020 at 6pm, due to alert level restrictions.

It will also be available later as a podcast.

 

There is near consensus in the emerging field of data ethics that processes and systems must be transparent and explainable to a wide range of stakeholders. 

In this talk, Professor Tim Dare from the Faculty of Arts at the University of Auckland will discuss why transparency and explainability have become central to data ethics and the reasons there are to question that centrality. Professor Dare will also discuss why we should be more concerned with reliability and with how automated systems compare, ethically, and with alternative ways of doing the tasks which might be done by automated systems.

See here for more details on the Gibbons Memorial Lecture Series.