Thought for the Week: Will A.I. do away with work?

On Saturday 24 February the Woking Debate considered Will AI do away with work?. It was attended by 20 people in the hall and 8 on Zoom. We were fortunate to get a great panel of speakers. William Baird, director of the UK branch of Pause AI started things off. Using a slide presentation he graphically showed how we are right to be scared. Every year AI is expanding its potential by three times. So if a human worker has an equal output to an AI this year, next year they will only be a third as productive as an AI and the year after only a ninth as productive unless they can also increase their productivity at a similar rate. A huge amount of money is being invested in training AIs to develop its skills. In 2014 AI became as good as us at reading other people’s hand writing, in 2016 it became as good as us at speech recognition and now has far exceeded us. AI can now create realistic voices and graphics. We were shown a short clip of a video that was very realistic and had been created in minutes by AI with a simple voice command. Carl Thomson, is an ex Conservative councillor. He acknowledged that AI has replaced work but questioned whether this was a threat or a transformation. Each new technology has replaced people and the key point is can we create better opportunities for people? Did YouTube replace the BBC as many people predicted or Spotify replace music companies? No they haven’t but they have had to adapt. AI does represent a significant advance in learning but it will not do away with work because we will create new fields of work with a demand for new specialists in cyber security, data analysis, interpretation of information and decision making. Some fields will be cut back but new fields will develop and there will always be a need for critical thinking and empathy. Teachers will still be vital. Customer service will still need people where bots reach their limits. Human control will remain absolutely necessary. You can ask Chat GPT for its code and it will provide it! Yet to do anything with it it will need people. The human eye will always be critical. Technology will only go as quickly as we let it. We need to regulate so we create the right balance. AI opens up tremendous opportunities for us. Paul Hoekstra is a member of the Green Party and a teacher in a performing arts college. In a thoughtful presentation he started by emphasising the AI potential was increasing every year and is seriously concerned that AI will mean his students will struggle to find work as AI can now create such realistic video images. It increases the danger of false videos being created of public figures and has made this much easier to do. No political parties are keeping pace with the changes in AI, including the Green Party. Will AI be used to do serious harm to people? Yet here are aspects of AI that are creating tremendous benefits such as lung cancer diagnosis where AIs have a better prediction rate than expert humans. Marilyn Taylor is a photographer who recently had an exhibition at the Lightbox whose display featured images using AI. She previously worked as a software programmer and so had some responsibility for people losing their jobs. AI now affects every aspect of photography allowing photographers to edit their photos. Yet there is a problem in blowing up the size of photos which can reveal its current limitations. It is now difficult to tell if photos have been altered. Questions and comments then came in. It sounds like AI is a psychopath. Carl thought this was an excellent point that he will use in future. The need for regulation was discussed and everyone thought this was imperative but difficult. It needs to be transparent and open and people need to be liable for the harm they cause. Paul talked about using AI to replace middle class jobs and skills. He proposed that a Universal Basic Income was the way to ameliate the suffering and allow people to follow their interests. Marilyn saw a danger in AIs learning from other AIs and how will we be able to know what is produced by humans. How do we get politicians to discuss these issues and care enough about them? We need to join unions and political parties and press these issues within them. We need more politicians with scientific backgrounds. Yet there is an Online Safety Act and Ofcom is a highly skilled regulator. AI created a video of Starmer shouting at his staff which was completely false and was circulated widely. Watermarks can be attached to video content so people can check out how genuine it is. The overall feeling was that AI can produce amazing benefits but we need to keep control of it and it needs to be properly regulated. With the benefits come dangers. To see the whole Debate of which this is only a short summary go on You Tube and look under Woking Debates. It will be online shortly. Keith Scott, member of the Quakers

Date: Tuesday 27th February 2024