After ChatGPT’s first birthday, has AI learned to walk?

In seasons past, the theatre of technology innovation saw fads often play the lead role, captivating audiences with their novelty before the spotlight shifted elsewhere.

There was the virtualisation hype, where computing boundaries were stretched. Then came the cloud era, which transformed the very fabric of IT infrastructure. More recently, there was the metaverse – arguably a one hit wonder, though most would argue it missed, eclipsed by the promise of AI.

This time last year, professionals returned to work to a whirlwind of ChatGPT talk, and it hasn’t stopped. The technology redefined the standards of AI and proved machines can ‘learn’ the complexities of human language and interaction (albeit inaccurately at times), which would enact automation that enhances both our productivity and creativity.

It made AI mainstream – a dinner table discussion. And now ChatGPT has had its first birthday. It begs the question, has AI learned to walk?

According to the CSIRO, 68% of Australian businesses have already implemented AI technologies, and a further 23% plan to implement some version of it in the next 12 months.

The technology isn’t a one-time show; it’s signing on for more seasons. But before organisations purchase their tickets, they need to recognise AI isn’t a fleeting thing – it needs to be nurtured and raised to maturity.

Nurturing AI for future success

Security threats, poor data quality, and privacy concerns over customer information are named the biggest concerns under the introduction of AI, according to the Australian Federal Government’s Export Finance Agency. The same report suggests most businesses will require at least four service providers to combat these risks and complex implementation.

But ultimately, businesses need to start with the basics, including not overlooking back-end systems and processes their AI feeds from.

The old expression “garbage in, garbage out” is extremely relevant to AI. If the available data is bad, the technology’s output is likely to be the same: When users find the AI is dishonest or simply broken, businesses can expect consequences.

Organisations need robust data to sustain their AI deployments, which includes knowing where data comes from and how it is being created. This is essential to trust the input that is feeding the system of intelligence.

This is where a context pipeline comes into the picture. With a data layer that is connected, clean, correct, secure, and in formats that can be synchronised and used effectively, a context pipeline feeds AI with data relevant to your enterprise. In a generative AI application, this would ensure the technology can deliver accurate and relevant answers to the questions you give it, and avoid many of the mishaps ChatGPT faces.

In essence, high-quality and complete data nourishes AI, but if the pantry is empty or lacks structure and organisation, AI systems are left underfed and fail to reach their promised outcomes.

This is why putting in the work upfront will pay off in the long run, particularly as businesses move toward trusting AI with bigger decisions.

Most businesses are still only comfortable targeting small productivity wins from AI, such as the creation of marketing content or enhancing internal communications and policy compliance.

The future of AI in decision-making and productivity

But we expect most human software decisions will become automated in the next five years. As processes mature and become more sophisticated, AI-driven disruption will become prominent.

We will be operating at a certain altitude in our jobs, and while human decision making will still be relevant, it will likely apply to more complex decisions.

In fact, we expect IT leaders could soon be leading teams of automated AI agents. Rather than waiting for a system to break, these intelligent agents will tell you when you are running hot, where the potential issues are, and be able to troubleshoot while you focus on more strategic and sophisticated work.

This will then begin to drive new services, product enhancements, or entirely new products, all of which amount to new revenue streams, and stronger mechanisms for personalised engagements with customers. The customer experience (CX) piece is gaining early prominence, with 77% of Australian CX leaders already experimenting with generative AI in their operations according to a report from Zendesk – that’s beyond the global average of 65%. Meanwhile, 70% plan to integrate it into numerous customer touch points over the next two years.

Another area companies will look to target are AI-driven recruitment, with algorithms helping to identify top talent and conducting preliminary checks and balances before smarter human decisions are made. Similarly, intelligent project management and autonomous procurement will disrupt traditional models.

AI stands poised, not as a fleeting fad, but as a transformative force that is already establishing a new order. Its ongoing success, however, will rest on a healthy recipe of government policy and data preparation.