AI as artists: Accessible artificial intelligence and the future of everything

It’s likely that you’ve come across at least some form of AI-generated art on the web. Most recently, OpenAI’s DALL-E and DALL-E 2 systems have been in the news. DALL-E/2 are AI systems that can generate realistic, surreal, technical or dreamy images, based on a natural language description. Really, they can generate whatever one’s mind is able to imagine, in shocking fidelity.

When Alan Turing conceptualised what an artificial intelligence machine could look like, he couldn’t have imagined the vagaries, ups and downs of AI development over the following decades. He’s remembered as one of the early thinkers in the space of artificial intelligence, and devised the eponymous “Turing test” in 1950, in which an AI must convince a battery of human interrogators that it is indeed human. No AI has succeeded in passing the strict test to date. However, Turing’s thoughts about machines that could learn through experience have come to fruition, and today, we see AI applied in many fields and indeed, in daily life.

All that speaks is not AI (but some is)

AI is an excellent marketing buzzword, and one that encompasses a variety of technologies that technically fall under the umbrella, but may not be the machine intelligence of science fiction. Rule-based systems – such as the early ELIZA program – may be able to pass off as AI, but based on where we are in tech today, it doesn't really cut it. Consider any chatbots you may have encountered on websites a few years ago (or even now). They’re scripted and coded to essentially serve up a menu of items based on a rudimentary set of rules. There is no learning, adapting and decision-making. Expert systems of the past also fall under the umbrella of AI, but are nowhere near the level of sophistication learning AI is today.

Modern AI is being applied to natural language processing (NLP), image recognition and complex applications such as autonomous cars. These can be considered good representations of where we are in artificial intelligence, since they combine various aspects of learning and data-driven decision-making. The DALL-E systems are an excellent example of this cutting-edge technology. By understanding and interpreting natural language input, and creating unreal images that look perfectly feasible, is a feat of artificial intelligence that is quite impressive, and certainly falls in the ambit of AI.

What this means for the future of work

The DALL-E systems, in some ways, are a seemingly ideal application for artificial intelligence – the machine interprets the human’s thoughts and desires, and brings them to fruition in an almost magical way. Consider the ship computers on the Star Trek sci-fi series. In essence, taking the variables of skill and rigour out of creation, the human mind is limited only by its imagination. There are, of course, many philosophical discussions to be had about the sanctity of work and its benefits.

OpenAI has kept the DALL-E systems closed to a select number of people, and expects to allow up to a million people to access it from their waitlist. The project is now moving from the research phase to the beta phase. Why a waitlist for such a seemingly wonderful use of technology? AI throws up many questions about its responsible use. With the DALL-E systems being so realistic and accurate in their renditions, it is likely that there may be opportunities for abuse, much as there are with “deepfake” videos – videos that appear to make people say things that they have not. AI is a tool that has the potential to exponentially magnify impact, which could be either good or bad. Humans designing AI systems will bring their own ethics into the design, which is problematic. To this end, the EU plans new legislation requiring AI to be responsible.

If AI is able to take all the skill out of a task, it also brings into question the future of skilled labour. What will happen to jobs that require years of practiced and acquired skill? We are not there yet for most tasks, but for many – especially ones that are more virtual than physical – we are rather close.

You may not know that it’s AI

Today, unlike in earlier times, AI is making a real business impact, and is present in daily life without being conspicuous. AI, in conjunction with behavioural sciences is providing insight to businesses like never before. In the wake of the pandemic, for instance, we had the “great resignation” where a large number of people exited the workforce. For some companies, AI has been able to see patterns that could otherwise be missed – employees with poorly-rated managers tended to quit more.

With the explosion of fintech and consumer-facing financial products through apps, AI is helping improve uptake by subtly changing messaging. For instance, AI is able to recommend credit limit increases for certain customers, but suspicion towards banks tends to dissuade customers from opting for added credit. AI helps modify the message – “why not keep reserves for a rainy day?” for instance. Coupled with design thinking, AI is making a measurable impact in finance.

Netflix and Google use AI to great effect for their online services. Most of Netflix’s recommendations for what you should watch next is powered by AI, looking at patterns in the shows you watch and suggesting others that may not be obvious. Google’s calendar alerts for flights is another great example – considering context (a flight to catch), traffic (real time data) to suggest a time to leave home (behavioural science) is AI making everyday impact.

AI has also quietly been making a difference where patterns and behaviours can throw up insights. In digital transformation initiatives to engage customers through omnichannel, in optimising supply chain for quicker fulfilment, for retail analytics to predict what customers will want to buy, and when to restock, and many other situations.

The future

With computing power and memory becoming more accessible and cloud computing becoming ubiquitous, AI is likely to continue its march. Quantum computing promises a giant leap forward in processing power, which should pay rich dividends for AI applications and make previously unthinkable tasks possible. Customer-facing industries such as retail, health, beauty will rely more on AI to track behaviour and tailor products and services to consumers. Healthcare will see AI making a significant impact, as we all move toward a quantified self in the age of smart devices. And there will most certainly be more wow moments.

Ajoy Singh

---

Ajoy Singh is the chief operating officer of Fractal.