Using a swearword in your Google search can stop that annoying AI overview from popping up. Some apps let you switch off their artificial intelligence.<\/p>\n
You can choose not to use ChatGPT, to avoid AI-enabled software, to refuse to talk to a chatbot. You can ignore Donald Trump posting deepfakes<\/a>, and dodge anything with Tilly the AI actor<\/a> in it.<\/p>\n As the use of AI spreads, so do concerns about its dangers, and resistance to its ubiquitousness.<\/p>\n Dr Kobi Leins \u2013 an AI management and governance expert \u2013 chooses to opt out when medical practitioners want to use AI.<\/p>\n She told a specialist she didn\u2019t want AI transcription software used for her child\u2019s appointment but was told it was necessary because the specialist was \u201ctime poor\u201d and if she did not want it used she would need to go somewhere else.<\/p>\n \u201cYou can\u2019t resist individually. There is also systemic resistance. The push from the industry to use these tools above and beyond where it makes sense [is so strong],\u201d she says.<\/p>\n Where is AI?<\/p>\n AI is spreading inexorably through digital systems.<\/p>\n It\u2019s embedded in applications such as ChatGPT, Google\u2019s AI overview, and Elon Musk\u2019s creation Grok, the super-Nazi chatbot<\/a>. Smartphones, social media and navigation devices are all using it.<\/p>\n It has also infiltrated customer service, the finance system, online dating apps and is being used to assess resumes and job applications, rental applications \u2013 even legal cases.<\/p>\n It\u2019s increasingly part of the healthcare system<\/a>, easing the administrative burden on doctors and helping to identify illnesses.<\/p>\n A global study from the University of Melbourne<\/a> released in April found half of Australians use AI on a regular or semi-regular basis, but only 36% trust it.<\/p>\n Prof Paul Salmon, the deputy director of the University of the Sunshine Coast\u2019s Centre for Human Factors and Sociotechnical Systems, says it\u2019s getting harder and harder to avoid.<\/p>\n \u201cIn work contexts, there is often pressure to engage with it,\u201d he says.<\/p>\n \u201cYou either feel like you\u2019re being left behind \u2013 or you\u2019re told you\u2019re being left behind.\u201d<\/p>\n Should I avoid using AI?<\/p>\n Privacy leakage, discrimination, false or misleading information, malicious use in scams and fraud, loss of human agency, lack of transparency and more are just some of the 1,600 risks in the Massachusetts Institute of Technology\u2019s AI Risk Database<\/a>.<\/p>\n It warns of the risk of AI \u201cpursuing its own goals in conflict with human goals or values\u201d and \u201cpossessing dangerous capabilities\u201d.<\/p>\n Greg Sadler, the chief executive officer of the Good Ancestors charity and coordinator of Australians for AI Safety, says he often refers to that database and, while AI can be useful, \u201cyou definitely don\u2019t want to use AI in the times where you don\u2019t trust its output, or you\u2019re worried about it having the information\u201d.<\/p>\n Aside from all those risks, AI has an energy cost. Google\u2019s emissions are up by more than 51%<\/a> thanks at least in part to the electricity consumption of datacentres that underpin its AI.<\/p>\n The International Energy Agency estimates that datacentres\u2019 electricity consumption could double from 2022 levels by 2026, while analysis shows they will be using 4.5% of global energy generation by 2030.<\/p>\n How can I avoid using AI?<\/p>\n AI overview has a \u201cprofanity trigger\u201d. If you ask Google \u201cWhat is AI?\u201d, its Gemini AI interface will deliver you a potted (and sometimes inaccurate) answer. It\u2019s functioning as an \u201canswer engine\u201d rather than a \u201csearch engine\u201d.<\/p>\n But if you ask \u201cWhat the fuck is AI?\u201d, you will be delivered straight search results, linking to other pages.<\/p>\n There are various browser extensions that can block AI sites, images and content.<\/p>\n You can circumvent some chatbots and speak to a human if you repeatedly say \u201cspeak to a human<\/a>\u201d, use the words \u201curgent\u201d and \u201cemergency\u201d, or, according to this report, \u201cblancmange<\/a>\u201d \u2013 a sweet dessert popular throughout Europe.<\/p>\n James Jin Kang, a senior lecturer in computer science at RMIT University Vietnam, writes in the Conversation<\/a> that to live entirely without it means \u201cstepping away from much of modern life\u201d.<\/p>\n \u201cWhy not just add a kill switch?\u201d he asks. The issue, he says, is that it is so embedded in our lives it is \u201cno longer something we can simply turn off\u201d.<\/p>\n \u201cSo as AI spreads further into every corner of our lives, we must urgently ask: will we still have the freedom to say no?<\/p>\n \u201cThe question isn\u2019t whether we can live with AI but whether we will still have the right to live without it before it\u2019s too late the break the spell.\u201d<\/p>\n What\u2019s the future of AI?<\/p>\n Governments around the world, including in Australia, are struggling to keep up with AI, what it means, what it promises and how to govern it.<\/p>\n As big tech companies seek access to material including journalism and books to train AI models, the federal government is under pressure<\/a> to reveal how it plans to regulate the technology.<\/p>\n