During this World AI Week, OneAdvanced is reflecting on the future of Artificial Intelligence (AI) and its impact on the workforce, exploring the concept of Co-Intelligence and how it is set to reshape the collaboration between humans and AI.
As AI continues to embed itself in the modern workplaces, there is a great deal of concern that goes with it. The discussion often defaults to ‘AI versus the workforce’, with fears of layoffs and hallucination. This worrying view is compelling but misses the true, emerging reality.
That new environment is something author Ethan Mollick, a professor at Wharton business school, calls ‘Co-Intelligence’. This is a new partnership where humans steer AI, blending the strengths of each to achieve better productivity, quality, and speed.
This blog post compliments our recent webinar ‘GenAI vs. the Workforce’ on the topic, a deep dive into AI’s effect on key business areas like cyber security and workforce productivity.
This blog will focus on the central topic, why AI still needs a cognitive helping hand and why the future lies in a human led partnership with AI.
The state of AI: beyond the hype
AI, particularly generative AI, has captured the imagination, and sometimes irritation, of both the tech world and the public. The launch of ChatGPT started what Gartner calls an ‘Innovation Trigger,' pushing AI into the mainstream.
However, as with all tech advancements, AI's journey has followed the Gartner Hype Cycle - peaking at inflated expectations before descending into the 'Trough of Disillusionment.' Early adopters are now questioning the return on investment, realising that while AI has potential, it is not a panacea.
Why is that?
Well, it is all to do with something humans are still much better at. Reasoning.
The reasoning gap: AI's achilles heel
Despite astonishing advancements, AI still struggles with tasks requiring reasoning and contextual understanding. Issues like 'hallucination,' where AI generates plausible but incorrect answers are widely known, but the real issue lies with common sense reasoning.
On X (formerly Twitter) you can easily find AI sceptics who trip up LLMs with questions a child could answer. A favourite of late has been ‘How many Rs in the word strawberry?’
ChatGPT fails again
LLMs are brilliant at generating reams of text but struggle when simple reasoning or mathematics is required. Generative AI would rather regurgitate training data, by predicting the next letter, than count, form a proper chain of thought or understand the actual physical world.
This failure to tackle simple cognitive tasks is what has bought users to the dreaded Trough of Disillusionment. Real breakthroughs are needed to create reasoning AI, but there is hope on the horizon, and strangely it also involves fruit.
Why are people asking Chat GPT about strawberries?
It has to do with what might be the next big breakthrough, Open AI’s unreleased model, codenamed ‘Strawberry’. AI watchers will know Strawberry has long been rumoured, a superior model with shiny new reasoning abilities, powered by an almost mythical research breakthrough named ‘Q*’. In a nutshell, Strawberry can allegedly reason and do maths and may be creating data for the next super model down the line, ‘Orion’.
Sam Altman’s teasing Strawberry tweet
Whether this is all the real deal, or so much fruit salad is yet to be revealed. The online hype around it has been something of a circus. For now, it is best to look at the real benefits that today’s AI is already bringing to the workplace.
AI in the workplace: the productivity booster
Incorporating AI into the workplace has already shown measurable benefits.
In collaboration with the Harvest Business School, Mollick and Wharton conducted a study with telling results. Two groups of consultants were given tasks, one of the groups was given access to AI. The consultants using AI tools completed 12.2% more tasks and finished them 25.1% faster than those without. Moreover, the quality of their output is over 40% higher, showcasing AI's potential as a productivity booster.
Early stats from Microsoft on the use of Copilot further back this up:
- 70% of Copilot users said they were more productive, and 68% said it improved the quality of their work.
- Overall, users were 29% faster in a series of tasks (searching, writing, and summarising).
- Users were able to get caught up on a missed meeting nearly 4x faster.
- 64% of users said Copilot helps them spend less time processing email.
- 85% of users said Copilot helps them get to a good first draft faster.
- 75% of users said Copilot “saves me time by finding whatever I need in my files.”
- 77% of users said once they used Copilot, they didn’t want to give it up.
Given AI's current capabilities and limitations, the future of work is not about choosing between AI or humans but about integrating the two. Mollick’s concept of 'Co-Intelligence' suggests a future where AI augments human abilities rather than replaces them.
For instance, Microsoft's Copilot can serve as an 'AI companion,' helping workers by automating repetitive tasks, summarising data, and managing projects, thereby reducing cognitive load and freeing up humans for more strategic and creative work. It simply is not capable of replacing workers yet, but it will disrupt processes and ultimately the way most people work.
Challenges ahead: the rise of Shadow AI
In fact, today’s AI is proving so useful, a new problem is emerging. Disruption, it seems, is coming to businesses, ready or not. As organisations consider the new reality, they must address the rising problem of ‘Shadow AI'—the use of unsanctioned AI tools by employees—which can lead to data leaks, compliance issues and legal problems.
Our recent webinar contains some alarming predictions from Forrester on the percentage of the workforce that will soon be bringing their own AI tools to the office. Clearly to realise safe Co-Intelligence, companies must invest in AI governance and security, ensuring that AI tools are integrated with official workflows.
The conclusion: embracing Co-Intelligence
In conclusion, the debate of AI versus the workforce misses the point. AI and humans are not yet adversaries, rather they are partners with humans in charge. By embracing Co-Intelligence, organisations can leverage AI's strengths - speed, efficiency, and data processing - while relying on human workers for their profound abilities in reasoning, creativity, and judgement.
This partnership can drive unprecedented levels of productivity and quality, but it requires thoughtful integration, continuous learning, and, crucially, managing creeping shadow AI.
For now, and likely for years to come, Co-Intelligence seems the sensible way forward. It frames a future where reasoning humans oversee powerful but flawed AI, each bringing the best from the other.
Interested in discussing this further? Get in touch today