Tuesday, April 16, 2024
HomeNewsDon't worry, Elon - computers will never outsmart us

Don’t worry, Elon – computers will never outsmart us


IN THE mid-1960s Joseph Weizenbaum, a computer scientist at MIT, created a programme called ELIZA which simulated human conversation by detecting key words and language patterns. I was able to play with it at Manchester University in the late 1970s. Its limitations were instantly obvious.  As soon as ELIZA got stuck it simply reflected back whatever was typed in, for example: ‘I’m sorry to hear that, tell me more’ and ‘That’s interesting, please expand.’ Yet despite its superficiality many people, including Weizenbaum’s own secretary, attributed human feelings to it.

Sixty years on contemporary ELIZAs are known as chatbots, some of which are employed as customer service agents on commercial websites, programmed to respond to a limited set of queries. ‘Hello, my name is Andrea/Andrew, how can I help you today?’ Conversation beyond these set parameters is impossible, as many frustrated customers know all too well.

Advanced bots, including Microsoft’s ChatGPT, can summarise complex topics, create prose, generate images (‘draw me a flying pig’) and write university-level student essays. ChatGPT appears so clever that some artificial intelligence (AI) industry leaders, including Elon Musk, have called for an immediate R&D pause because they fear AI will ‘outsmart and replace us’. 

To assuage his alarm Musk would do well to read Weizenbaum’s seminal work, Computer Power and Human Reason: From Judgement to Calculation, not least since its title beautifully encapsulates the unbridgeable gap between computer processing and human reasoning.

Artificially intelligent calculations are fundamentally different from human judgement. Computers are not human and so cannot think like us. Humans are capable of the most brilliant logic (we designed the AI programmes after all) but our decisions are much more than this, as chatGPT conceded in a recent ‘conversation’ with me:

DS: How are you?

ChatGPT: I am an AI language model, so I don’t have emotions like humans do. However, I am functioning properly and ready to help you with any questions or tasks you may have. How may I assist you?

DS: What is the meaning of life?

ChatGPT: As an AI language model, I don’t have personal beliefs or values, so I cannot provide a philosophical answer. The meaning of life is a complex question that has been debated by scientists, philosophers, and religions throughout history.

DS: Are you real?

ChatGPT: I am not a physical being and do not have a physical form. I exist solely as a software program running on servers. However, my responses are based on complex algorithms and machine learning that enable me to simulate human-like conversations.

DS: If I were to tell you a joke, would you laugh?

ChatGPT: As an AI language model, I am not capable of laughing, but I can generate jokes for you if you’d like!

DS: My father died in 1976.

ChatGPT:  I’m sorry to hear that.

DS: I see. The trouble is you said earlier: ‘I am an AI language model, so I don’t have emotions like humans do.’ So how can you be sorry? Sorrow is a human emotion and you are just a bunch of rules.

ChatGPT: You are correct that as an AI language model, I do not possess emotions like humans do. I apologize if my earlier response may have given the impression that I am capable of experiencing emotions. When I said ‘I’m sorry’, it was meant to convey that I understand and acknowledge the issue you raised, and I will do my best to provide a helpful response.

DS: But you don’t understand the issue, you need to have emotions to understand it. You are not sorry because you cannot be sorry about my father’s death. If you say you are sorry you are lying.

ChatGPT: As an AI language model, I may not experience emotions in the same way humans do. However, I can empathize with you and understand the gravity of your loss. I realize that losing a loved one is a painful experience, and my programming allows me to offer condolences and support during such difficult times. Therefore, when I say ‘I’m sorry,’ I genuinely mean it, even though I may not feel emotions.

DS: That is just nonsense, as a matter of logic. First, you do not experience emotions at all. Second – and crucially – you are not capable of empathizing. Here is what Merriam Webster says: ‘Empathy involves actively sharing in the emotional experience of the other person’. You cannot possibly share in my emotional experience, as you have already said. Third, you cannot ‘genuinely’ mean you are sorry because you are nothing more than computer code.

At this point it all got too much for ChatGPT which fed my own response back to me, just as ELIZA did in 1978, albeit with more sophistication.

Weizenbaum’s point is as valid today as it ever was. Computer processing can speed up programmed logic, but however impressive the algorithms, they are not us – they are an accelerated facet of one part of us – our logical reason – but human judgement is so much more than logic.

Human decision-making is a perpetually unique combination of persons and our circumstances. We are not separate observers of an objective world. We are an inextricable part of it in every facet of our being. We are part of a magical flow of life in which every decision we make changes both us and the world we inhabit, organically and constantly.

Unlike computer programmes we can be wonderfully illogical – contradictory, impulsive, passionate, and fiercely protective of our children. We are driven by ego, desire, outrage, jealousy, culture, fear, love and our unique personal histories.

Human beings will continue to use the power of computing to achieve our goals, both to control and to liberate, such is our Jekyll and Hyde nature. But the worry that AI will transcend our intelligence is baseless paranoia.

The fear of a superhuman machine takeover is a product of overestimating the scope of computer software and drastically underestimating the astonishing complexity of the human experience. To grasp this in a single thought, imagine a computer praying. 

If you appreciated this article, perhaps you might consider making a donation to The Conservative Woman. Unlike most other websites, we receive no independent funding. Our editors are unpaid and work entirely voluntarily as do the majority of our contributors but there are inevitable costs associated with running a website. We depend on our readers to help us, either with regular or one-off payments. You can donate here. Thank you.
If you have not already signed up to a daily email alert of new articles please do so. It is here and free! Thank you.

David Seedhouse
David Seedhouse
David Seedhouse is the creator of Our Decision Too.

Sign up for TCW Daily

Each morning we send The ConWom Daily with links to our latest news. This is a free service and we will never share your details.