AI is the hottest topic in technology, dominating headlines over the past eighteen months. This panel, chaired by iResearch Services’ Gurpreet Purewal, garnered a lot of attention.
Evolution or revolution?
Gurpreet began the session by asking the first panelist, Philippe Roussiere, Global Lead for Innovation and AI at Accenture Research, if AI was a revolution or just a natural progression.
“You’re asking a Frenchman if it’s a revolution?” responded Philippe, to plenty of laughter from the audience. He spoke about how some people view this technology as a utility, like electricity. From his experience, he stated that to gain value from it, it is vital to have a lot of data as a foundation, to then build the capabilities, prepare your workforce and have the talent in place.
From his research, Philippe has discovered that AI transformation is happening faster than digital transformation.
“I think it’ll change the way we work,” said Cindy Anderson, the Global Executive of Thought Leadership Engagement and Eminence at the IBM Institute for Business Value, describing it as more of a revolution than evolution. She predicted we will have digital workers and AI sitting next to us in many areas “really soon,” and described it as a “dramatic change.”
Cindy’s colleague, Anthony Marshall, Senior Research Director of Thought Leadership at the IBM Institute for Business Value added he cannot remember such a technology that has impacted CEOs to such a degree. He remarked that there is huge pressure on CEOs to react.
The business response to AI in the workplace
Gurpreet asked the panel about the trends in business at the C-Suite level when it comes to AI.
“Definitely bullish on Generative AI,” was Philippe’s response, stating that there is a belief that it is transformative and ready to change the workforce. He cited the possibility of Generative AI impacting 40% of working hours. The big goals, he said, were to maximize value and minimize risk. On risk, he mentioned negative outcomes including AI hallucinations, bias, misinformation and data leakage.
We were told by Philippe that 70% of the C-Suite “have gone through some form of education” on deploying AI for business advantages.
Cindy pointed out that there is a lack of clarity on what Generative AI means, and that executives tended to be overconfident on the issue. She pointed out there are some cautious voices raising concern about the data AI is being trained on, such as intellectual property (IP) issues around that data and cyber security issues.
These issues being raised prompted Gurpreet to ask the panel: “Why do you think the challenges have changed when the attitudes haven’t adjusted?”
Anthony observed that in terms of priorities, there seemed to be a lot of ambition, but also a lot of fear – especially surrounding litigation. He also pointed out there is a lack of understanding of the technology.
Gurpreet asked the panel: “How does AI help inform your thought leadership?”
“I was reading the nice piece of research that talks about the Leading Lights squad,” responded Philippe, referring to iResearch Services’ report on how thought leadership affects business leaders. “So, the Data Cruncher and the Ignitra (innovation-focused) persona in there. That actually is my purpose in Accenture Research, to define those use cases where we want to scale or accelerate AI and Gen AI in particular.”
Philippe then pointed out how Generative AI was particularly good in workflow areas like ideation and testing ideas from survey data. He conceded some fact-checking and validation exercises would be necessary.
Cindy was keen to point out that in AI, things are changing so fast. She said, “Since about July we’ve been producing a report for CEOs on Generative AI every two weeks!” These IBM reports have touched on ‘deep dive’ subjects, such as talent, customer and employee experience, and tech spend.
Anthony has noticed “an increase in the consumption of thought leadership over the past year,” stating that a typical CEO now spends two and a half hours a week on thought leadership. He and his team believed Generative AI would be used in the production of thought eadership, and he did not think this would negate expertise.
Augmenting AI rather than depending on it
These insights prompted Gurpreet to talk about a possible consequence – organizations falling into a ‘disparity’ of Generative AI. He acknowledged that this technology could help with ideas but is only as good as the information it is based on. He asked the panel on the best way to augment AI to supercharge ideas, rather than being completely reliant on it.
Philippe was keen to point out that AI technology is moving fast in terms of language apprehension, discovering nuance and its ability to respond to more human-like prompts.
When it comes to using new data sources, Philippe pointed out that Generative AI can ingest data in the form of vector databases. Usually made for Large Language Models (LLMs), these are collections of data stored as mathematical representations of objects or data points in a multi-dimensional space, where each dimension is related to an attribute or feature.
He described: “It’s a way to make the AI smarter with the data that you trust and data that your researchers would normally turn to, in a human way. Except this time the machine has the power to analyze much more of it.”
His next point was that humans are there to apply data ethics. This included critical thinking, fact-checking and avoiding the spurious correlations machines can make.
Cindy warned us that her observation would be a little controversial. She believes it is “really hard” to produce a good prompt for Generative AI. This led her to conclude that if someone is good at being a prompt engineer, they will succeed at producing compelling thought leadership.
“I don’t know when,” Cindy conceded, “maybe in six months, maybe a year, maybe tomorrow, I don’t know!”
This is where an audience member wanted to engage. He pointed out there were a number of talented prompt engineers with online communities on Discord and X (formally known as Twitter). He also stated that focus should be on the principles of good prompts, rather than just prompts themselves.
Cindy was pleased to hear this, and said she knew of talented prompt engineers who IBM’s research showed were commanding “massively high salaries.”
Don’t believe the hype?
On the subject of predictions, another audience member raised a question on the Hype Cycle and said he believed that Generative AI’s hype has peaked at the ‘Inflated Expectations’ step. The audience member wanted to know if the panel has an estimate as to when it will hit the ‘Trough of Disillusionment’?”
Philippe pointed out that Generative AI’s progress is hitting a point where robots can be on a factory floor and learn how to assist in the work process so that it is more than just a professional sitting at a computer entering prompts. He speculated that the general knowledge of LLMs may hit a limit for being applied to real-world cases. He said industry-specific models in the consumer goods sector could well be the next wave and if it happens fast, there would be no ‘Trough of Disillusionment’.
It was also pointed out that the cost of computing is coming down rapidly, with Philippe noting that Nvidia’s new generation of AI chips is 1,000 times faster than the previous one. Gurpreet mentioned a potential issue of this technical advantage; power consumption may have a negative impact on a company’s sustainability goals.
Anthony said, “Right now, Generative AI is like a very smart undergraduate that is totally non-numeric. So, Generative AI, it’s a language model, it’s not statistical, its application to thought leadership is very specific. It’s very good at taking a huge amount of data.”
He raised the point that there are many unknown areas where it could be taken, but to be wary of its conclusions, just like any other type of analytical technology.
Risks of embracing AI
The panel took another question from the audience: “What’s the risk level of embracing AI to create thought leadership?”
Anthony believed it to be low and made the point that good thought leadership is always practical and conclusions should not be obvious.
Another audience member asked, tongue-in-cheek, if we should all be preparing for a four-day week and more leisure time.
Philippe was positive in his answer, stating that AI could liberate the researcher to get more ‘white space’ and avoid the drudgery of research.
One of the major criticisms of AI, whether it could take a huge swathe of jobs, was the theme of one audience member’s question, who asked how CEOs could allay the fears of their workforce.
Cindy cited Mike Read, the CEO of WPP, whose views on this subject were published in IBM’s most recent action guide. She said anybody who is in a knowledge worker capacity, or a creative capacity should be worried. Since the public availability of ChatGPT in November, she has told her team to study new AI breakthroughs every week. Her summary is: “Generative AI isn’t going to replace people, but people who use Generative AI are going to replace people who don’t.” She ended on an upbeat point, that AI would also create new job opportunities.
Philippe touched on that latter point, advising the audience that they should get to know their new AI ‘colleague’, learn prompt engineering, understand scaling, and build on their strengths.
“Fear of technological innovation has been around since the Industrial Revolution,” remarked Anthony. He believed that while some CEOs are lacking education in this area, thought leadership can have a role to play in shining a light on the future of AI.
The future workforce will blend humans and technology
Finally, Yogesh Shah, CEO of iResearch Services, asked the panel what working life will be like years from now. Cindy replied, “I expect my job to be completely different a year from now.” She admitted she does not know exactly how, but predicted she would be managing a group of humans and AI assistants.
Philippe was positive about new roles, such as ‘AI Ethicist’ or ‘Machine Learning Therapist’.
Anthony responded: “I’m hoping to be happily retired, waving to my chatbot from a beach in Thailand, who’s doing my job,” which garnered much laughter. However, he did end on a serious point, recalling that in his examination of customer service, AI has the capacity to radically change the nature of employees’ skill sets. The future of AI is a double-edged sword, one that holds the promise of unparalleled innovation, and the potential for unforeseen challenges. It is a testament to the connection and collaboration between human intellect and machine intelligence that this concluding paragraph has been generated by ChatGPT and carefully edited by humans. A partnership that can illuminate the path forward, while never replacing the vital role of human wisdom and ethical guidance.
Interested in watching the To AI, or not to AI? session on-demand? Click here.Back to Blogs