They say that reputation is everything. It was only a few years ago that Steven Spielberg admitted he regretted filming Jaws as it fuelled fear and harmed shark conservation efforts. Some would argue that AI finds itself in a similar position. For every Johnny 5 that demonstrates that humans and robots can work together, there’s a Terminator or Ultron that shows them attempting to destroy the world. This leaves AI fighting to improve its public perception - something that isn’t helped by misaligned communication, poor strategic implementation and crippling oversight.
Propeller Group’s latest panel at Tech Show London 2025 sought to address how to shift the narrative around the great technological disruptor. AI's PR Challenge: Strategies to Overcome the Myths and Misconceptions, moderated by our MADtech Practice Director Alex Humphries-French, featured several industry leaders:
- Tom Laranjo, CEO of Mediaplus UK / Behave
- Supriya Dev-Purkaystha, Head of Media and Ad Tech Solutions, UK at Microsoft
- Lauren Hendry Parsons, Vice President of Communications & Advocacy at ExpressVPN
AI won’t take our jobs - but could infringe on our values
Tom Laranjo opened the session by expanding on a popular notion that although AI isn’t here to take people’s jobs, the people who use AI will. He outlined it won’t be enough for workers to loosely use AI and expect job security; businesses must invest in AI training for their personnel. “We're in a moment of radical transformation, not optimisation. We’re going to need budgets to re-educate, re-skill ourselves and our workforces”.
In other words, businesses can’t afford to be casual with investing in AI transformation - a sentiment echoed by Supriya Dev-Purkaystha, who added perspective on how AI benefits human productivity and reduces decision fatigue. “We’re not creating AI to replace humans; these tools are meant to enable us to do more. AI is supporting us when making important decisions, ensuring we make the right decision.”
However, fears around AI extend beyond job security. Its arrival on the global stage has been accompanied by a host of ethical concerns. For Lauren Hendry Parsons, the key to addressing these challenges is via collective action. “AI can be ethical - but it needs to be a choice that it happens. It can’t be achieved alone. Not by companies, not by regulators and not by individuals. It needs to happen on a systemic level.”
Bottom line: the existing regulatory bodies and systems we have in place aren’t equipped to resolve issues around environmental sustainability or the reality of a global internet. Given the need for regulatory alignment on AI ethics, the onus is on businesses to grasp the opportunity with both hands, do the hard yards and shape ethical AI in line with their own values.
AI chatbot’s are only as intelligent as we configure them…for now
The conversation then turned to the role humans have in detracting or improving AI’s logical and interpersonal intelligence.
It’s commonly asserted that AI exhibits greater intelligence than humans - but Supriya doesn’t think it’s that simple. “Gen AI is dependent on the data and training that we give it. This means its intelligence is dependent on us. As such, it’s up to us to keep AI in check and ensure it’s evolving at the right pace, in the right ways. And for this reason it’s important that the teams building and developing are diverse in their skills and lived experiences.”
Lauren approached the myth from a different angle, fearing over-reliance on AI will make humans dumber. She argued that personal agency is so important in regards to our usage of AI; formulating a good prompt indicates the capabilities are all there with us humans. “Let AI shoulder the heavy lifting for the tasks that it’s actually effective at, such as huge data cuts. When it comes to strategic thinking, it’s important for us to keep flexing our muscles.”
Continuing this school of thought, Lauren spoke on the role that training and data have on AI’s inherent biases. More importantly, she spoke on the impact that neglecting interpersonal intelligence has had. “Humans are biased. Humans train AI models. Therefore, the models are biased. Even worse, they can perpetuate that bias. That’s why its strengths lie in assigning it to tasks where there’s no interpretation required, such as synthesising, analysing and extracting the relevant information from rafts of data.”
“I can’t use AI - it’s too expensive!”
Businesses’ hesitancy to adopt AI into their way of working is well documented, and Tom suggests this gap between awareness and implementation will do latecomers no favours.
“There’s a lot of expense and effort wasted trying to build AI solutions and applications which, in the not-too-distant future, will be delivered by larger providers. Rather than focussing all of your resources on something which is soon going to be replaced, the focus should be on investing in the skills required to operate these new solutions. It’s all about continual learning.”
Supriya then elaborated on Microsoft’s efforts to combat misconceptions around AI’s expensive nature, and how it was democratising these tools. “We put Copilot into Bing, a platform that’s free and available to all.. The decision to implement Gen AI into Bing made sense, as search engines were beginning to struggle to keep up with users’ increasingly complex searches. As an organisation our mission is to empower everybody on the planet to do more, Copilot does just that”
In regards to AI’s reputation problem, Tom acknowledged that both businesses and the general public will be slow to jump from ‘A to I’ without a concerted effort to improve awareness. This means that the priority has to be building confidence and proficiency in using these tools.
“Change is generally regarded as scary. Getting past this negativity requires organisations educating their audiences. More than that, they need to listen - taking people’s concerns on board allows you to show how you’re going to shape the way that AI is interacted with.”
Building on this, Lauren argued that the time for action is now, particularly given that people are already bringing their own devices into work. “Support and accommodation is paramount. People’s quality of work will improve if they’re backed up with these tools. The key is to properly integrate it within your IT ecosystem and ensure you have the right checks in place.”
Final words of wisdom
The final minutes of the panel saw each speaker discuss how the wider industry can improve AI adoption. Supriya kicked us off, with her thoughts focussing on psychological safety. “People need to be able to give honest feedback, especially when it’s not working. Organisations can build trust by adopting agile methodology; and encourage fast failure. It’s OK to come up with your hypotheses and use cases, test them and be ok if they fail”
For Lauren, transparency and accountability are the order of the day. “Be honest about how you’re using the technology and the data. Own up to it - make it clear that when they engage with AI, this is what they’re getting. If things don’t go to plan, be accountable. Acknowledge what went wrong and identify how you’re going to fix it and ensure it doesn’t happen again. That’s the key to winning trust.”
Tom’s closing thoughts emphasised the importance of humility and vulnerability. “We’re all learning a lot, and it’s probably fair to say that we’re all faking it before we make it right now. This is where the vulnerability comes in; there’s nothing wrong with saying that we’re experimenting and exploring simultaneously.
We’re approaching a moment of profound transformation. Many themes are colliding at once - biotechnology, improved microsensors, quantum computing. It’s essential that we meet that moment with profound humility and develop a comprehensive understanding of this change before rushing head-first.”