The recent flurry, or rather blizzard, of announcements of new variants of generative AI has caused a storm of hype and fear. OpenAI’s ChatGPT already seemed like a game changer, but now this week’s new version, GPT-4, is another step forward. GPT-4 can generate enough text to write a book, encode in all computer languages, and most remarkably, “understand” images.
If your mind isn’t boggled by the potential of this, then you haven’t been paying attention. I have spent the last five years researching how artificial intelligence has changed journalism around the world. I have seen how it can supercharge the media to collect, create and distribute content in much more efficient and effective ways. It is already the “next wave” of technological change. Now, generative AI has moved potential progress up a speed or two.
But wait. This isn’t much of a breakthrough for “aware” AI. Robots are not coming to replace us. However, these Long Language Models (LLMs), like ChatGPT, are an accelerator that operates at such scale and speed that they can seem to do whatever you ask them to do. And the more we use them and provide them with data and questions, the faster they learn to predict outcomes.
A million startups already claim to use this secret sauce to create new products that will revolutionize everything from legal administration to stock trading to gaming to medical diagnostics. A lot of this is marketing foam. As with all technological advances, there is always an exaggerated cycle and unintended good and bad consequences. But I’ve seen enough to know that it’s going to change our lives. Just think what these tools could do when used by creative people in fashion or architecture, for example.
Artificial intelligence such as machine learning, automation or natural language processing is already part of our world. For example, when you search online, it uses machine learning-based algorithms trained on large data sets to give you what you’re looking for. Now the pace of change is accelerating. In 2021 alone, global private corporate investment in AI has doubled, and I expect advances in generative AI to double it again.
Now take a breath. I don’t recommend anyone using ChatGPT or GPT-4 to build anything at this time, at least not something that will be used without human verification to make sure it’s accurate, reliable, efficient, and doesn’t do any harm. AI isn’t about fully automating end-to-end content production – it’s about augmentation to give professionals and creatives the tools to work faster, freeing them up to spend more time on what humans do best.
We know that there are some real additional risks in using generative AI. He has “hallucinations” where he makes things up. Sometimes creates harmful content. And it will surely be used to spread misinformation or to invade privacy. People have already used it to create new ways to hack into computers, for example. You might want to use it to create a wonderful new video game, but what if some arch-villain uses it to create a deadly virus?
We know about those risks because we can see their flaws when we test these prototypes that technology companies have made available to the public. You can have a lot of fun making him write poems or songs or create surreal images. Ask him a direct question and you’ll usually get a no-nonsense, confident answer. Ask him a stupid or complex question, and he will fight. Many techies and journalists have had fun testing it to destruction and making it respond in strange and disturbing ways. AI boffins will be delighted because all of this helps refine their programming. They are conducting their experimentation partly in public.
We also know the risks because OpenAI itself has listed them on its “system card” which explains the new powers and dangers of this technology, and how it has tried to improve them with each new iteration. Who ultimately decides what risks are acceptable or what we should do about them is a moot question.
It is too late to put this technology “back in the box”. It has too much potential to help humans deal with the global challenges we face. It is vital that we have an open debate about the ethical, economic, political and social impact of all forms of AI. I hope our politicians quickly educate themselves on this rapidly growing technology better than in the past, and that we all become more AI literate. But ultimately my main hope is that we take the time and effort to think carefully about the best ways it can be used in a positive way. You don’t have to believe the hype to have some hope.
