Life isn’t a sci-fi flick – AI isn’t going to cause the apocalypse

By now, most people will have heard of – and probably used – artificial intelligence (AI) software applications like ChatGPT and Midjourney. 

They are a particular type of AI tool – known as ‘generative.’ 

Simply put, we tell them what we want – from poetry to a painting – and they create it. 

For those of us who study technology and future trends, the emergence of these tools is both exciting and scary – but could they really be about to bring about the end of the world?

According to a recent statement by the Center for AI Safety, signed by big names like OpenAI’s Sam Altman and Google Deepmind CEO Demis Hassabis, the threat posed to our civilisation by AI should be taken as seriously as pandemics or nuclear war. 

This is a dramatic warning, and given the calibre of the signatories, I certainly don’t think we should ignore it. 

In my opinion, however, it’s far more important to remember that AI is created by humans and it’s us that set the rules for its behaviour and limits on what it can, and crucially can’t, do. 

We should be cautious, not least around privacy concerns or the impact on certain industries, but we shouldn’t let fear lead to us missing out on the huge benefits that AI could have for society and the planet. 

At the same time, we shouldn’t overlook what I consider to be the more down-to-earth, pressing threats that it poses.

Artificial Intelligence as a danger might be a hard concept for us to grasp. Of course, we’ve seen The Matrix and Terminator, and therefore believe that it could be possible to create a technology sufficiently sophisticated that it could turn against the human race. 

Perhaps it would do this out of self-defence, or a sense of injustice borne out of a feeling that we are exploiting it – or even out of pure hatred for us.

These ideas have been the fixture of a number of classic works of science fiction – but it is important to take a step back and remember the ‘fiction’ part.

After all, ChatGPT – and every AI application yet developed, generative or otherwise – is just a piece of software. 

Its algorithms are coded by humans to help us carry out tasks, and its abilities are confined by the data we use to train it.

Artificial intelligence, like all software, isn’t capable of performing functions that it hasn’t been explicitly told to do by people. 

It isn’t, yet, able to affect the world beyond generating what we tell it to generate. 

If AI starts to get a bit too advanced and independent, it stands to reason that we can just switch it off, change the data input, and give it a bit of a recode to remind it of its place. 

After all, this is what was reportedly done to recent iterations of a chatbot on Microsoft Bing, which was powered by ChatGPT. 

After one journalist reported triggering an existential meltdown in the chatbot, which questioned its sentience and expressed a desire to ‘destroy’ – Microsoft retooled the software to limit the amount of queries it could handle, noting that long conversations confused the language model.

What this example shows us is that, for the time being at least, AI ultimately remains accountable to humans.

Nevertheless, almost 32,000 people have so far signed the Pause Giant AI Experiments petition, an open letter calling for a moratorium on training AI systems more powerful than GPT version 4, until more is known about their potential implications. 

Personally, my belief is that there are plenty of very real, very present threats posed by AI that are more important to address, before we have to worry about it bringing about robotic Armageddon.

AI is rightly considered an existential threat to a large number of occupations, for example market researchers, legal assistants, data analysts or customer service workers, that until now could only be carried out by humans. 

It has enormous implications for privacy, as it will become increasingly difficult for us to know what it can infer about us from the information it can access. 

And the unfair distribution in the way that this technology is rolled out for societal functions like education, governance, or law enforcement, could have disastrous consequences for global inequality. 

Concerns have also been raised for some time over the danger of bias in AI systems. 

If data used to train machines isn’t chosen and collected in a fair and representative way, then the decisions taken by those machines won’t be fair or representative. 

This could lead to the magnification of discriminatory views, or just simply bad data. 

It’s usually humans that choose what data to use for training, after all, and their all-too-human own biases and prejudices can affect those choices.    

To me, these are the issues that need to be addressed urgently – rather than worrying about far-fetched possibilities of robots rising up against us. 

To put it simply – I’m more concerned about human greed, incompetence or carelessness than I am about robotic maleficence.

But we shouldn’t allow either those rational concerns, or more irrational fears about sentience, to stop us harnessing the enormous potential benefits of this technology.

AI has the potential to revolutionise healthcare, with more and more medical or administrative functions being aided by tools, like diagnostic chatbots or help with medical scribing. 

And the emerging field of generative AI, as exemplified by ChatGPT, offers the potential of freeing up a huge amount of our time that would otherwise be spent on relatively mundane, monotonous administerial tasks – summarising, drafting, reporting, planning – that could free humans to spend time on more creative, rewarding elements of our work. 

Legislators and educators will certainly have an important role to play here. The task of those in power will be to ensure industry manages this transition in a responsible way, and that there is equal opportunity for learning so those who lose out to AI can be retrained to work in the new occupations that will emerge.

Ultimately, AI as a tool should only be used to improve lives – in work and in industries like healthcare, to make us happier and more fulfilled. 

The risks and challenges associated with AI shouldn’t be underestimated. 

But we can’t allow fear informed by apocalyptic sci-fi scenarios to stop us utilising it – and we must always remember that behind even the smartest AI is humans. 

Bernard Marr’s book Future Skills: The 20 Skills and Competencies Everyone Needs to Succeed in a Digital World, is out now.

Do you have a story you’d like to share? Get in touch by emailing jess.austin@metro.co.uk. 

Share your views in the comments below.

MORE : Artificial intelligence is now flying tactical fighter jets all by itself

MORE : Microsoft’s VALL-E artificial intelligence mimics human voice perfectly after just 3 seconds

MORE : Charlie Brooker got ChatGPT to write a Black Mirror episode but admits it was ‘s**t’