Red herrings are things that distract us from what is really important. A bit like bringing up the upcoming movie Superman vs Voldemort during a conversation that is hitting a lull*.

When talking about artificial intelligence (AI), it’s easy to get caught up in abstract debates relating to “intelligence” and consciousness. While these topics are fascinating, I believe they can be red herrings that can distract us from the real-world implications of AI as it stands, and as it’s likely to stand as it continues progressing.

For me, the important thing is whether, when, how, and to what extent, AI is going to impact our lives. Abstract definitions of intelligence and consciousness are often beside the point.

What is “intelligence”, really?

We already have AI. We have machines that can perform tasks that resemble, or require, intelligence, including forms of intelligence that seem to be uniquely human.

Electronic calculators are a great example. Pocket calculators have been widely available since the 1970s. They can add, subtract, multiply, divide, and find the square root of numbers more quickly and accurately than any human being. Scientific, graphing, and financial calculators can perform even more sophisticated tasks.

Microsoft Word has been able to check the spelling and grammar of writing for decades, and often picks up errors that someone editing the same piece of writing will miss. Services like Grammarly provide advanced versions of these capabilities, including giving you a sense of the the tone of your writing and how it might come across to others.

Google Photos can identify photographs from based on a very basic prompt. I can get it to search for images with specific people or pets in them. Or I can ask it to look for specific items in my photo album, like candles or toilets. (Or candles and toilets…)

I can ask DALL-E 2 to create an image based on a short text description of an image I want, and I’ll be blown away but what it creates. I can ask ChatGPT to answer questions I wouldn’t reasonably expect any normal human to be able to answer. I can use it as a sounding board to guide decisions, or “think out loud”. It can guide me to do things with Excel or Python that I wouldn’t otherwise be able to do, or would take me hours or days to learn.

These are all forms of artificial intelligence. Whether or not one or more of these systems understands what they are doing, have any degree of self-awareness, can independently improve themselves, or meet some arbitrary threshold of “creativity” doesn’t change the fact that they are excellent at certain types of tasks.

Whether tools like this are narrowly intelligent (calculators, Google photos) or more broadly intelligent (ChatGPT), the underlying reality is that these different forms of AI are are tools that can augment our own intelligence and capabilities.

AGIs, creativity, recursive self-improvement, and so on

There are some situations where it might be important whether an AI meets some definition of AGI or “artificial general intelligence”. Perhaps it will require some unique type of creativity, or deeper understanding of the world, or an ability to engage in “recursive self-improvement”.

Adding some arbitrary measure that AI needs to meet before it is worth thinking about or discussing in detail, is weird to me. Frankly, I’m not too concerned about whether an AI meets some definition of “strong AI” or “general intelligence”. I am more concerned about the impact of AI, in whatever form it comes.

Consciousness

On that note, another red herring is whether an AI can achieve, will achieve, or has achieved, consciousness. AI systems don’t have to be conscious to be useful or destructive.

Personally, I think the issue that arises from one or more AI systems becoming conscious are the ethical implications we will (or should) have to face. If we create something with consciousness, and it’s capable of suffering, then there are legitimate questions that need to be asked regarding the moral status we afford them. (On this note, I’m not very optimistic that we’ll make the “right” decisions, based the moral status we give to many animals that are part of our food supply. But I digress.)

* Superman versus Voldemort

On the topic of digressions, have you heard that a new movie is coming out called Superman versus Voldemort?

It’s not. And it will never come out*. But it’s a factoid I bring out when a conversation needs to be redirected. Something of a red herring, you could say.

* Unless AI continues advancing, in which case we ’ll probably be able to watch movies based on our specific prompts. In which case, there almost certainly will be a version of a movie called Superman versus Voldemort .

More concrete questions

The main reason I’ve written this article is that AI seems to be one of the few topics where the people I chat with seem to talk abstractly rather than concretely. Which is funny, because in most cases I can probably be accused of wanting to talk too abstractly.

But I want to be involved in more concrete discussions about how it’s going to impact us. Whether it meets certain definitions of “intelligence” or whether it’s consciousness are, to a large extent, beyond the point. There are concrete implications of these technologies, and they’re speeding up regardless.

For example:

  • What are characteristics of employment roles that are at risk, and that are likely to be safe?

  • How are people and organisations likely to use AI to be more efficient and effective at certain tasks?

  • If some tasks can be performed more efficiently and effectively, what are the next-best things that we can do with our time and energy? (Let’s face it: there’s virtually no end to the things that could be done. Almost no one has nothing on their “to do” list. The issue is prioritising what needs to be done.)

  • How can AI make things worse, and what can we do to prevent, minimise, and mitigate these possibilities?

  • How can AI make the world better in the short-, medium-, and long-term? And how can we use it in our own lives? For our own benefit, for the benefit of our loved ones, and for the world in general?

AI systems are tools. Whether we use these tools well remains to be seen. How can we use the improved capabilities that these tools afford us, to build our lives, and our world, even more in line with our values and priorities?