AI Won’t Take Your Job, but It May Drastically Change It
Introduction
There are many speculations that AI will soon take our jobs. It makes sense that there would be; my feed is filled with news about technological advancements and we see it everywhere in our daily lives. I now find it completely normal for Amazon to summarize product reviews based on my preferences, or to see an AI-automated search engine surfacing key points when researching online. AI lives alongside us, and it will become more and more seamlessly integrated into everything we do.
One thing is clear though: AI gives us helpful information, but it does not complete the end goal we're looking to achieve. For example, when on my honeymoon I asked ChatGPT, "What are exciting things to do when spending a day in Melbourne?". I entered enough information for the AI to learn my preferences and it produced a great plan, accurate in location and timings. But, it couldn't go that one step further in its helpfulness and actually book the day out, purchasing museum tickets or a rooftop brunch on my behalf. Clearly, the need for human processing is still required as AI does not have the sophistication yet to do the actual 'action' needed.
In our professional life, things are similar. AI does speed up our work, creating document drafts, meeting notes and summaries, and seriously helps us save time and remain accurate - but it also still requires a human to perform a task. We are not yet into the realm of fully autonomous agents, able to mimic the human and interact on their behalf. AI is incredible and it's always evolving, but currently it's surfaced and utilized as a chatbot style of interaction, made more useful by layering automation triggers for next steps, though not yet capable to do that of its human counterparts.
Advancements
That said, there are continuous advancements taking place for tools we love. Brand new functionality, from AI Builder (my favourite, and what I consider to be the “OG” AI tool for Power Platform) to new and improved Agents. Check out April Dunnam’s video on AI Builder’s recent features to learn more.
“Github Copilot helped developers code up to 55% faster, and 85% of developers had increased confidence.”
We also have GitHub Copilot for free now, which is an incredible tool. It works like a knowledgeable teammate and pair programmer, allowing developers to produce better code, faster. Specifically, Github Copilot helped developers code up to 55% faster, and has resulted in 85% of developers having higher confidence in coding (Github, 2024). Though, this tool only helps if the human themselves has a good understanding of code to begin with, and can validate the Copilot outputs and rectify any inaccuracies which might be present in the code.
Furthermore, ground breaking news such as the chip produced by Google, Willow, is said to significantly reduce errors as it scales up; a major breakthrough in quantum error correction. Willow can perform a computation in under five minutes that would take a supercomputer 10 septillion years, demonstrating its potential for solving complex problems beyond the reach of classical computers. This achievement marks a significant step towards building commercially relevant quantum computers that can revolutionize fields like medicine, energy, and AI. This is extremely impressive and could be a significant addition to the effort to cure disease or to explore space, as long as humans know how to use it!
With all this incredible progress in AI, albeit helpful, time saving and sometimes ground breaking, it still does not replace the human in any scenario. So, what does the future look like?
A few days ago, Satya Nadella sat down for an interview where he talked clearly about his vision for Copilot. I highly recommend you watch the video.
In this interview, Satya says that copilot is our organizing layer, to help humans be more productive. He talks about how Agents may come to replace core business applications, because people want more AI-enabled BizApps to help our daily working lives, so we are expecting to see more day-to-day processes being orchestrated via Agents.
Copilot agents take on mundane and repetitive tasks, freeing up human time to focus on more strategic and creative endeavours. By automating routine activities, such as data entry, scheduling, and report generation, these agents help us accomplish more in less time.
Trustworthy AI
With these clear operational improvements enabling people to innovate more, and the speed of which AI is emerging within everything we do, at this juncture we really ought to be thinking less about if AI will take our jobs, and more about working alongside AI. Considering how we verify AI's reliability in producing the outputs we intended, accurately and without bias. The critical component to consider right now is Trustworthy AI.
Whether we use AI tools within our organization, or we develop new products, we MUST take responsibility for these solutions. This technology is generative, thereby generating unique responses to masses of (potentially sensitive) data, providing users with outputs that differ with each unique prompt. As part of of this change we must ensure that users are aware it’s their responsibility to verify AI outputs, and the organization's responsibility to red team (test) AI for non-bias outputs which are accurate, reliable and transparent in how the agents arrived at their answers.
It isn't just me encouraging this, it's the law: EU AI Act First Regulation on Artificial Intelligence.
Don't panic, you don't have to start from scratch - Microsoft have created a guide for this around 'Trustworthy AI': Microsoft Trustworthy AI: Unlocking human potential starts with trust - The Official Microsoft Blog
Conclusion
To sum up, AI helps humans speed up their work and allows them time for innovation. No, AI will not take your job, but it will drastically change it (for the better!). As we progress forwards, we cannot know precisely where this fast-moving technological landscape will take us. But, we can surmise that as agents become more sophisticated in their ability to perform actions, autonomously, then humans will need to increase their sophistication in observing its outputs and doing more with their time to advance key initiatives.
For now, learn how to use AI tools; acquire new skills to create, use and govern AI workloads and double down on Trustworthy AI, and you'll be just fine.